Example data sets and computer code for the book chapter titled "Missing Data in the Analysis of Multilevel and Dependent Data" submitted for publication in the second edition of "Dependent Data in Social Science Research" (Stemmler et al., 2015). This repository includes the computer code (".R") and the data sets from both example analyses (Examples 1 and 2). The data sets are available in two file formats (binary ".rda" for use in R; plain-text ".dat").

The data sets contain simulated data from 23,376 (Example 1) and 23,072 (Example 2) individuals from 2,000 groups on four variables:

ID = group identifier (1-2000) x = numeric (Level 1) y = numeric (Level 1) w = binary (Level 2)

In all data sets, missing values are coded as "NA".

Presented are propensity score mean-squared-error and standardised ration of propensity score mean-squared error.

This dataset contains simulated datasets, empirical data, and R scripts described in the paper: “Li, Q. and Kou, X. (2021) WiBB: An integrated method for quantifying the relative importance of predictive variables. Ecography (DOI: 10.1111/ecog.05651)”.

A fundamental goal of scientific research is to identify the underlying variables that govern crucial processes of a system. Here we proposed a new index, WiBB, which integrates the merits of several existing methods: a model-weighting method from information theory (Wi), a standardized regression coefficient method measured by ß* (B), and bootstrap resampling technique (B). We applied the WiBB in simulated datasets with known correlation structures, for both linear models (LM) and generalized linear models (GLM), to evaluate its performance. We also applied two other methods, relative sum of wight (SWi), and standardized beta (ß*), to evaluate their performance in comparison with the WiBB method on ranking predictor importances under various scenarios. We also applied it to an empirical dataset in a plant genus Mimulus to select bioclimatic predictors of species’ presence across the landscape. Results in the simulated datasets showed that the WiBB method outperformed the ß* and SWi methods in scenarios with small and large sample sizes, respectively, and that the bootstrap resampling technique significantly improved the discriminant ability. When testing WiBB in the empirical dataset with GLM, it sensibly identified four important predictors with high credibility out of six candidates in modeling geographical distributions of 71 Mimulus species. This integrated index has great advantages in evaluating predictor importance and hence reducing the dimensionality of data, without losing interpretive power. The simplicity of calculation of the new metric over more sophisticated statistical procedures, makes it a handy method in the statistical toolbox.

Methods To simulate independent datasets (size = 1000), we adopted Galipaud et al.’s approach (2014) with custom modifications of the data.simulation function, which used the multiple normal distribution function rmvnorm in R package mvtnorm(v1.0-5, Genz et al. 2016). Each dataset was simulated with a preset correlation structure between a response variable (y) and four predictors(x1, x2, x3, x4). The first three (genuine) predictors were set to be strongly, moderately, and weakly correlated with the response variable, respectively (denoted by large, medium, small Pearson correlation coefficients, r), while the correlation between the response and the last (spurious) predictor was set to be zero. We simulated datasets with three levels of differences of correlation coefficients of consecutive predictors, where ∆r = 0.1, 0.2, 0.3, respectively. These three levels of ∆r resulted in three correlation structures between the response and four predictors: (0.3, 0.2, 0.1, 0.0), (0.6, 0.4, 0.2, 0.0), and (0.8, 0.6, 0.3, 0.0), respectively. We repeated the simulation procedure 200 times for each of three preset correlation structures (600 datasets in total), for LM fitting later. For GLM fitting, we modified the simulation procedures with additional steps, in which we converted the continuous response into binary data O (e.g., occurrence data having 0 for absence and 1 for presence). We tested the WiBB method, along with two other methods, relative sum of wight (SWi), and standardized beta (ß*), to evaluate the ability to correctly rank predictor importances under various scenarios. The empirical dataset of 71 Mimulus species was collected by their occurrence coordinates and correponding values extracted from climatic layers from WorldClim dataset (www.worldclim.org), and we applied the WiBB method to infer important predictors for their geographical distributions.

This archive contains code and data for reproducing the analysis for “Replication Data for Revisiting ‘The Rise and Decline’ in a Population of Peer Production Projects”. Depending on what you hope to do with the data you probabbly do not want to download all of the files. Depending on your computation resources you may not be able to run all stages of the analysis. The code for all stages of the analysis, including typesetting the manuscript and running the analysis, is in code.tar. If you only want to run the final analysis or to play with datasets used in the analysis of the paper, you want intermediate_data.7z or the uncompressed tab and csv files. The data files are created in a four-stage process. The first stage uses the program “wikiq” to parse mediawiki xml dumps and create tsv files that have edit data for each wiki. The second stage generates all.edits.RDS file which combines these tsvs into a dataset of edits from all the wikis. This file is expensive to generate and at 1.5GB is pretty big. The third stage builds smaller intermediate files that contain the analytical variables from these tsv files. The fourth stage uses the intermediate files to generate smaller RDS files that contain the results. Finally, knitr and latex typeset the manuscript. A stage will only run if the outputs from the previous stages do not exist. So if the intermediate files exist they will not be regenerated. Only the final analysis will run. The exception is that stage 4, fitting models and generating plots, always runs. If you only want to replicate from the second stage onward, you want wikiq_tsvs.7z. If you want to replicate everything, you want wikia_mediawiki_xml_dumps.7z.001 wikia_mediawiki_xml_dumps.7z.002, and wikia_mediawiki_xml_dumps.7z.003. These instructions work backwards from building the manuscript using knitr, loading the datasets, running the analysis, to building the intermediate datasets. Building the manuscript using knitr This requires working latex, latexmk, and knitr installations. Depending on your operating system you might install these packages in different ways. On Debian Linux you can run apt install r-cran-knitr latexmk texlive-latex-extra. Alternatively, you can upload the necessary files to a project on Overleaf.com. Download code.tar. This has everything you need to typeset the manuscript. Unpack the tar archive. On a unix system this can be done by running tar xf code.tar. Navigate to code/paper_source. Install R dependencies. In R. run install.packages(c("data.table","scales","ggplot2","lubridate","texreg")) On a unix system you should be able to run make to build the manuscript generalizable_wiki.pdf. Otherwise you should try uploading all of the files (including the tables, figure, and knitr folders) to a new project on Overleaf.com. Loading intermediate datasets The intermediate datasets are found in the intermediate_data.7z archive. They can be extracted on a unix system using the command 7z x intermediate_data.7z. The files are 95MB uncompressed. These are RDS (R data set) files and can be loaded in R using the readRDS. For example newcomer.ds <- readRDS("newcomers.RDS"). If you wish to work with these datasets using a tool other than R, you might prefer to work with the .tab files. Running the analysis Fitting the models may not work on machines with less than 32GB of RAM. If you have trouble, you may find the functions in lib-01-sample-datasets.R useful to create stratified samples of data for fitting models. See line 89 of 02_model_newcomer_survival.R for an example. Download code.tar and intermediate_data.7z to your working folder and extract both archives. On a unix system this can be done with the command tar xf code.tar && 7z x intermediate_data.7z. Install R dependencies. install.packages(c("data.table","ggplot2","urltools","texreg","optimx","lme4","bootstrap","scales","effects","lubridate","devtools","roxygen2")). On a unix system you can simply run regen.all.sh to fit the models, build the plots and create the RDS files. Generating datasets Building the intermediate files The intermediate files are generated from all.edits.RDS. This process requires about 20GB of memory. Download all.edits.RDS, userroles_data.7z,selected.wikis.csv, and code.tar. Unpack code.tar and userroles_data.7z. On a unix system this can be done using tar xf code.tar && 7z x userroles_data.7z. Install R dependencies. In R run install.packages(c("data.table","ggplot2","urltools","texreg","optimx","lme4","bootstrap","scales","effects","lubridate","devtools","roxygen2")). Run 01_build_datasets.R. Building all.edits.RDS The intermediate RDS files used in the analysis are created from all.edits.RDS. To replicate building all.edits.RDS, you only need to run 01_build_datasets.R when the int... Visit https://dataone.org/datasets/sha256%3Acfa4980c107154267d8eb6dc0753ed0fde655a73a062c0c2f5af33f237da3437 for complete metadata about this dataset.

The PATH Study was launched in 2011 to inform the Food and Drug Administration's regulatory activities under the Family Smoking Prevention and Tobacco Control Act (TCA). The PATH Study is a collaboration between the National Institute on Drug Abuse (NIDA), National Institutes of Health (NIH), and the Center for Tobacco Products (CTP), Food and Drug Administration (FDA). The study sampled over 150,000 mailing addresses across the United States to create a national sample of people who use or do not use tobacco. 45,971 adults and youth constitute the first (baseline) wave, Wave 1, of data collected by this longitudinal cohort study. These 45,971 adults and youth along with 7,207 "shadow youth" (youth ages 9 to 11 sampled at Wave 1) make up the 53,178 participants that constitute the Wave 1 Cohort. Respondents are asked to complete an interview at each follow-up wave. Youth who turn 18 by the current wave of data collection are considered "aged-up adults" and are invited to complete the Adult Interview. Additionally, "shadow youth" are considered "aged-up youth" upon turning 12 years old, when they are asked to complete an interview after parental consent. At Wave 4, a probability sample of 14,098 adults, youth, and shadow youth ages 10 to 11 was selected from the civilian, noninstitutionalized population (CNP) at the time of Wave 4. This sample was recruited from residential addresses not selected for Wave 1 in the same sampled Primary Sampling Unit (PSU)s and segments using similar within-household sampling procedures. This "replenishment sample" was combined for estimation and analysis purposes with Wave 4 adult and youth respondents from the Wave 1 Cohort who were in the CNP at the time of Wave 4. This combined set of Wave 4 participants, 52,731 participants in total, forms the Wave 4 Cohort. At Wave 7, a probability sample of 14,863 adults, youth, and shadow youth ages 9 to 11 was selected from the CNP at the time of Wave 7. This sample was recruited from residential addresses not selected for Wave 1 or Wave 4 in the same sampled PSUs and segments using similar within-household sampling procedures. This "second replenishment sample" was combined for estimation and analysis purposes with the Wave 7 adult and youth respondents from the Wave 4 Cohorts who were at least age 15 and in the CNP at the time of Wave 7. This combined set of Wave 7 participants, 46,169 participants in total, forms the Wave 7 Cohort. Please refer to the Restricted-Use Files User Guide that provides further details about children designated as "shadow youth" and the formation of the Wave 1, Wave 4, and Wave 7 Cohorts. Dataset 0002 (DS0002) contains the data from the State Design Data. This file contains 7 variables and 82,139 cases. The state identifier in the State Design file reflects the participant's state of residence at the time of selection and recruitment for the PATH Study. Dataset 1011 (DS1011) contains the data from the Wave 1 Adult Questionnaire. This data file contains 2,021 variables and 32,320 cases. Each of the cases represents a single, completed interview. Dataset 1012 (DS1012) contains the data from the Wave 1 Youth and Parent Questionnaire. This file contains 1,431 variables and 13,651 cases. Dataset 1411 (DS1411) contains the Wave 1 State Identifier data for Adults and has 5 variables and 32,320 cases. Dataset 1412 (DS1412) contains the Wave 1 State Identifier data for Youth (and Parents) and has 5 variables and 13,651 cases. The same 5 variables are in each State Identifier dataset, including PERSONID for linking the State Identifier to the questionnaire and biomarker data and 3 variables designating the state (state Federal Information Processing System (FIPS), state abbreviation, and full name of the state). The State Identifier values in these datasets represent participants' state of residence at the time of Wave 1, which is also their state of residence at the time of recruitment. Dataset 1611 (DS1611) contains the Tobacco Universal Product Code (UPC) data from Wave 1. This data file contains 32 variables and 8,601 cases. This file contains UPC values on the packages of tobacco products used or in the possession of adult respondents at the time of Wave 1. The UPC values can be used to identify and validate the specific products used by respondents and augment the analyses of the characteristics of tobacco products used

Source:

Mariusz Šapczyński, Cracow University of Economics, Poland, lapczynm '@' uek.krakow.pl Sylwester Białowąs, Poznan University of Economics and Business, Poland, sylwester.bialowas '@' ue.poznan.pl

Data Set Information:

The dataset contains information on clickstream from online store offering clothing for pregnant women. Data are from five months of 2008 and include, among others, product category, location of the photo on the page, country of origin of the IP address and product price in US dollars.

Attribute Information:

The dataset contains 14 variables described in a separate file (See 'Data set description')

Relevant Papers:

N/A

Citation Request:

If you use this dataset, please cite:

Šapczyński M., Białowąs S. (2013) Discovering Patterns of Users' Behaviour in an E-shop - Comparison of Consumer Buying Behaviours in Poland and Other European Countries, “Studia Ekonomiczne†, nr 151, “La société de l'information : perspective européenne et globale : les usages et les risques d'Internet pour les citoyens et les consommateurs†, p. 144-153

Data description ìe-shop clothing 2008î

Variables:

1. YEAR (2008)

========================================================

2. MONTH -> from April (4) to August (8)

========================================================

3. DAY -> day number of the month

========================================================

4. ORDER -> sequence of clicks during one session

========================================================

5. COUNTRY -> variable indicating the country of origin of the IP address with the

following categories:

1-Australia 2-Austria 3-Belgium 4-British Virgin Islands 5-Cayman Islands 6-Christmas Island 7-Croatia 8-Cyprus 9-Czech Republic 10-Denmark 11-Estonia 12-unidentified 13-Faroe Islands 14-Finland 15-France 16-Germany 17-Greece 18-Hungary 19-Iceland 20-India 21-Ireland 22-Italy 23-Latvia 24-Lithuania 25-Luxembourg 26-Mexico 27-Netherlands 28-Norway 29-Poland 30-Portugal 31-Romania 32-Russia 33-San Marino 34-Slovakia 35-Slovenia 36-Spain 37-Sweden 38-Switzerland 39-Ukraine 40-United Arab Emirates 41-United Kingdom 42-USA 43-biz (.biz) 44-com (.com) 45-int (.int) 46-net (.net) 47-org (*.org)

========================================================

6. SESSION ID -> variable indicating session id (short record)

========================================================

7. PAGE 1 (MAIN CATEGORY) -> concerns the main product category:

1-trousers 2-skirts 3-blouses 4-sale

========================================================

8. PAGE 2 (CLOTHING MODEL) -> contains information about the code for each product

(217 products)

========================================================

9. COLOUR -> colour of product

1-beige 2-black 3-blue 4-brown 5-burgundy 6-gray 7-green 8-navy blue 9-of many colors 10-olive 11-pink 12-red 13-violet 14-white

========================================================

10. LOCATION -> photo location on the page, the screen has been divided into six parts:

1-top left 2-top in the middle 3-top right 4-bottom left 5-bottom in the middle 6-bottom right

========================================================

11. MODEL PHOTOGRAPHY -> variable with two categories:

1-en face 2-profile

========================================================

12. PRICE -> price in US dollars

========================================================

13. PRICE 2 -> variable informing whether the price of a particular product is higher than

the average price for the entire product category

1-yes 2-no

========================================================

14. PAGE -> page number within the e-store website (from 1 to 5)

++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Abstract

China Living Standards Survey (LSS) consists of one household survey and one community (village) survey, conducted in Hebei and Liaoning Provinces (northern and northeast China) in July 1995 and July 1997 respectively. Five villages from each three sample counties of each province were selected (six were selected in Liaoyang County of Liaoning Province because of administrative area change). About 880 farm households were selected from total thirty-one sample villages for the household survey. The same thirty-one villages formed the samples of community survey. This document provides information on the content of different questionnaires, the survey design and implementation, data processing activities, and the different available data sets.

Geographic coverage

Regional

Analysis unit

Households

Kind of data

Sample survey data [ssd]

Sampling procedure

The China LSS sample is not a rigorous random sample drawn from a well-defined population. Instead it is only a rough approximation of the rural population in Hebei and Liaoning provinces in North-eastern China. The reason for this is that part of the motivation for the survey was to compare the current conditions with conditions that existed in Hebei and Liaoning in the 1930's. Because of this, three counties in Hebei and three counties in Liaoning were selected as "primary sampling units" because data had been collected from those six counties by the Japanese occupation government in the 1930's. Within each of these six counties (xian) five villages (cun) were selected, for an overall total of 30 villages (in fact, an administrative change in one village led to 31 villages being selected). In each county a "main village" was selected that was in fact a village that had been surveyed in the 1930s. Because of the interest in these villages 50 households were selected from each of these six villages (one for each of the six counties). In addition, four other villages were selected in each county. These other villages were not drawn randomly but were selected so as to "represent" variation within the county. Within each of these villages 20 households were selected for interviews. Thus, the intended sample size was 780 households, 130 from each county. Unlike county and village selection, the selection of households within each village was done according to standard sample selection procedures. In each village, a list of all households in the village was obtained from village leaders. An "interval" was calculated as the number of the households in the village divided by the number of households desired for the sample (50 for main villages and 20 for other villages). For the list of households, a random number was drawn between 1 and the interval number. This was used as a starting point. The interval was then added to this number to get a second number, then the interval was added to this second number to get a third number, and so on. The set of numbers produced were the numbers used to select the households, in terms of their order on the list. In fact, the number of households in the sample is 785, as opposed to 780. Most of this difference is due to a village in which 24 households were interviewed, as opposed to the goal of 20 households

Mode of data collection

Face-to-face [f2f]

Cleaning operations

(a) DATA ENTRY All responses obtained from the household interviews were recorded in the household questionnaires. These were then entered into the computer, in the field, using data entry programs written in BASIC. The data produced by the data entry program were in the form of household files, i.e. one data file for all of the data in one household/community questionnaire. Thus, for the household there were about 880 data files. These data files were processed at the University of Toronto and the World Bank to produce datasets in statistical software formats, each of which contained information for all households for a subset of variables. The subset of variables chosen corresponded to data entry screens, so these files are hereafter referred to as "screen files". For the household survey component 66 data files were created. Members of the survey team checked and corrected data by checking the questionnaires for original recorded information. We would like to emphasize that correction here refers to checking questionnaires, in case of errors in skip patterns, incorrect values, or outlying values, and changing values if and only if data in the computer were different from those in the questionnaires. The personnel in charge of data preparation were given specific instructions not to change data even if values in the questionnaires were clearly incorrect. We have no reason to believe that these instructions were not followed, and every reason to believe that the data resulting from these checks and corrections are accurate and of the highest quality possible.

(b) DATA EDITING The screen files were then brought to World Bank headquarters in Washington, D.C. and uploaded to a mainframe computer, where they were converted to "standard" LSMS formats by merging datasets to produce separate datasets for each section with variable names corresponding to the questionnaires. In some cases, this has meant a single dataset for a section, while in others it has meant retaining "screen" datasets with just the variable names changed. Linking Parts of the Household Survey Each household has a unique identification number which is contained in the variable HID. Values for this variable range from 10101 to 60520. The first number is the code for the six counties in which data were collected, the second and third digits are for the villages within each county. Finally, the last two digits of HID contain the household number within the village. Data for households from different parts of the survey can be merged by using the HID variable which appears in each dataset of the household survey. To link information for an individual use should be made of both the household identification number, HID, and the person identification number, PID. A child in the household can be linked to the parents, if the parents are household members, through the parents' id codes in Section 01B. For parents who are not in the household, information is collected on the parent's schooling, main occupation and whether he/she is currently alive. Household members can be linked with their non-resident children through the parents' id codes in Section 01C. Linking the Household to the Community Data The community data have a somewhat different set of identifying variables than the household data. Each community dataset has four identifying variables: province (code 7 for Hebei and code 8 for Liaoning); county (six two digit codes, of which the first digit represents province and the second digit represents the three counties in each province); township (3 digit code, first digit is county, second digit is county and third digit is township); and village (4 digit code, first digit is county, second digit is county, third digit is township, and third fourth digit is village). Constructed Data Set Researchers at the World Bank and the University of Toronto have created a data set with information on annual household expenditures, region codes, etc. This constructed data set is made available for general use with the understanding that the description below is the only documentation that will be provided. Any manipulation of the data requires assumptions to be made and, as much as possible, those assumptions are explained below. Except where noted, the data set has been created using only the original (raw) data sets. A researcher could construct a somewhat different data set by incorporating different assumptions. Aggregate Expenditure, TOTEXP. The dataset TOTEXP contains variables for total household annual expenditures (for the year 1994) and variables for the different components of total household expenditures: food expenditures, non-food expenditures, use value of consumer durables, etc. These, along with the algorithm used to calculate household expenditures are detailed in Appendix D. The dataset also contains the variable HID, which can be used to match this dataset to the household level data set. Note that all of the expenditure variables are totals for the household. That is, they are not in per capita terms. Researchers will have to divide these variables by household size to get per capita numbers. The household size variable is included in the data set.

Abstract

Survey based Harmonized Indicators (SHIP) files are harmonized data files from household surveys that are conducted by countries in Africa. To ensure the quality and transparency of the data, it is critical to document the procedures of compiling consumption aggregation and other indicators so that the results can be duplicated with ease. This process enables consistency and continuity that make temporal and cross-country comparisons consistent and more reliable.

Four harmonized data files are prepared for each survey to generate a set of harmonized variables that have the same variable names. Invariably, in each survey, questions are asked in a slightly different way, which poses challenges on consistent definition of harmonized variables. The harmonized household survey data present the best available variables with harmonized definitions, but not identical variables. The four harmonized data files are

a) Individual level file (Labor force indicators in a separate file): This file has information on basic characteristics of individuals such as age and sex, literacy, education, health, anthropometry and child survival. b) Labor force file: This file has information on labor force including employment/unemployment, earnings, sectors of employment, etc. c) Household level file: This file has information on household expenditure, household head characteristics (age and sex, level of education, employment), housing amenities, assets, and access to infrastructure and services. d) Household Expenditure file: This file has consumption/expenditure aggregates by consumption groups according to Purpose (COICOP) of Household Consumption of the UN.

Geographic coverage

National

Analysis unit

• Individual level for datasets with suffix _I and _L
• Household level for datasets with suffix _H and _E

Universe

The survey covered all de jure household members (usual residents).

Kind of data

Sample survey data [ssd]

Sampling procedure

SAMPLE DESIGN FOR ROUND 4 OF THE GLSS A nationally representative sample of households was selected in order to achieve the survey objectives.

Sample Frame For the purposes of this survey the list of the 1984 population census Enumeration Areas (EAs) with population and household information was used as the sampling frame. The primary sampling units were the 1984 EAs with the secondary units being the households in the EAs. This frame, though quite old, was considered inadequate, it being the best available at the time. Indeed, this frame was used for the earlier rounds of the GLSS.

Stratification In order to increase precision and reliability of the estimates, the technique of stratification was employed in the sample design, using geographical factors, ecological zones and location of residence as the main controls. Specifically, the EAs were first stratified according to the three ecological zones namely; Coastal, Forest and Savannah, and then within each zone further stratification was done based on the size of the locality into rural or urban.

SAMPLE SELECTION EAs A two-stage sample was selected for the survey. At the first stage, 300 EAs were selected using systematic sampling with probability proportional to size method (PPS) where the size measure is the 1984 number of households in the EA. This was achieved by ordering the list of EAs with their sizes according to the strata. The size column was then cumulated, and with a random start and a fixed interval the sample EAs were selected.

It was observed that some of the selected EAs had grown in size over time and therefore needed segmentation. In this connection, such EAs were divided into approximately equal parts, each segment constituting about 200 households. Only one segment was then randomly selected for listing of the households.

Households At the second stage, a fixed number of 20 households was systematically selected from each selected EA to give a total of 6,000 households. Additional 5 households were selected as reserve to replace missing households. Equal number of households was selected from each EA in order to reflect the labour force focus of the survey.

NOTE: The above sample selection procedure deviated slightly from that used for the earlier rounds of the GLSS, as such the sample is not self-weighting. This is because, 1. given the long period between 1984 and the GLSS 4 fieldwork the number of households in the various EAs are likely to have grown at different rates. 2. the listing exercise was not properly done as some of the selected EAs were not listed completely. Moreover, it was noted that the segmentation done for larger EAs during the listing was a bit arbitrary.

Mode of data collection

Face-to-face [f2f]

This paper explores a unique dataset of all the SET ratings provided by students of one university in Poland at the end of the winter semester of the 2020/2021 academic year. The SET questionnaire used by this university is presented in Appendix 1. The dataset is unique for several reasons. It covers all SET surveys filled by students in all fields and levels of study offered by the university. In the period analysed, the university was entirely in the online regime amid the Covid-19 pandemic. While the expected learning outcomes formally have not been changed, the online mode of study could have affected the grading policy and could have implications for some of the studied SET biases. This Covid-19 effect is captured by econometric models and discussed in the paper. The average SET scores were matched with the characteristics of the teacher for degree, seniority, gender, and SET scores in the past six semesters; the course characteristics for time of day, day of the week, course type, course breadth, class duration, and class size; the attributes of the SET survey responses as the percentage of students providing SET feedback; and the grades of the course for the mean, standard deviation, and percentage failed. Data on course grades are also available for the previous six semesters. This rich dataset allows many of the biases reported in the literature to be tested for and new hypotheses to be formulated, as presented in the introduction section. The unit of observation or the single row in the data set is identified by three parameters: teacher unique id (j), course unique id (k) and the question number in the SET questionnaire (n ϵ {1, 2, 3, 4, 5, 6, 7, 8, 9} ). It means that for each pair (j,k), we have nine rows, one for each SET survey question, or sometimes less when students did not answer one of the SET questions at all. For example, the dependent variable SET_score_avg(j,k,n) for the triplet (j=Calculus, k=John Smith, n=2) is calculated as the average of all Likert-scale answers to question nr 2 in the SET survey distributed to all students that took the Calculus course taught by John Smith. The data set has 8,015 such observations or rows. The full list of variables or columns in the data set included in the analysis is presented in the attached filesection. Their description refers to the triplet (teacher id = j, course id = k, question number = n). When the last value of the triplet (n) is dropped, it means that the variable takes the same values for all n ϵ {1, 2, 3, 4, 5, 6, 7, 8, 9}.Two attachments:- word file with variables description- Rdata file with the data set (for R language).Appendix 1. Appendix 1. The SET questionnaire was used for this paper. Evaluation survey of the teaching staff of [university name] Please, complete the following evaluation form, which aims to assess the lecturer’s performance. Only one answer should be indicated for each question. The answers are coded in the following way: 5- I strongly agree; 4- I agree; 3- Neutral; 2- I don’t agree; 1- I strongly don’t agree. Questions 1 2 3 4 5 I learnt a lot during the course. ○ ○ ○ ○ ○ I think that the knowledge acquired during the course is very useful. ○ ○ ○ ○ ○ The professor used activities to make the class more engaging. ○ ○ ○ ○ ○ If it was possible, I would enroll for the course conducted by this lecturer again. ○ ○ ○ ○ ○ The classes started on time. ○ ○ ○ ○ ○ The lecturer always used time efficiently. ○ ○ ○ ○ ○ The lecturer delivered the class content in an understandable and efficient way. ○ ○ ○ ○ ○ The lecturer was available when we had doubts. ○ ○ ○ ○ ○ The lecturer treated all students equally regardless of their race, background and ethnicity. ○ ○

Migration flows are derived from the relationship between the location of current residence in the American Community Survey (ACS) sample and the responses given to the migration question "Where did you live 1 year ago?". There are flow statistics (moved in, moved out, and net moved) between county or minor civil division (MCD) of residence and county, MCD, or world region of residence 1 year ago. Estimates for MCDs are only available for the 12 strong-MCD states, where the MCDs have the same government functions as incorporated places. Migration flows between metropolitan statistical areas are available starting with the 2009-2013 5-year ACS dataset. Flow statistics are available by three or four variables for each dataset starting with the 2006-2010 5-year ACS datasets. The variables change for each dataset and do not repeat in overlapping datasets. In addition to the flow estimates, there are supplemental statistics files that contain migration/geographical mobility estimates (e.g., nonmovers, moved to a different state, moved from abroad) for each county, MCD, or metro area.

Two different machine-learning methods (classification and regression trees (CART) and random forests (RF)) were applied to the artificially generated data set 4, which comprises two groups with sizes of n = 1000 cases and d = 20 variables. The results represent the medians of test performance measures from 100-fold cross-validation runs using random splits of the data set into training data (2/3 of the data set) and test data (1/3 of the data set). In addition, a negative control data set was created by permutating the variables from the training data set, with the expectation that the machine learning algorithms should not perform group assignment better than chance when trained with such data; otherwise, there could be overfitting involved.

Abstract (en): The purpose of this project was to measure and estimate the distribution of personal income and related economic factors in both rural and urban areas of the People's Republic of China. The principal investigators based their definition of income on cash payments and on a broad range of additional components. Data were collected through a series of questionnaire-based interviews conducted in rural and urban areas at the end of 2002. There are ten separate datasets. The first four datasets were derived from the urban questionnaire. The first contains data about individuals living in urban areas. The second contains data about urban households. The third contains individual-level economic variables copied from the initial urban interview form. The fourth contains household-level economic variables copied from the initial urban interview form. The fifth dataset contains village-level data, which was obtained by interviewing village leaders. The sixth contains data about individuals living in rural areas. The seventh contains data about rural households, as well as most of the data from a social network questionnaire which was presented to rural households. The eighth contains the rest of the data from the social network questionnaire and is specifically about the activities of rural school-age children. The ninth dataset contains data about individuals who have migrated from rural to urban areas, and the tenth dataset contains data about rural-urban migrant households. Dataset 1 contains 151 variables and 20,632 cases (individual urban household members). Dataset 2 contains 88 variables and 6,835 cases (urban households). Dataset 3 contains 44 variables and 27,818 cases, at least 6,835 of which are empty cases used to separate households in the file. The remaining cases from dataset 3 match those in dataset 1. Dataset 4 contains 212 variables and 6,835 cases, which match those in dataset 2. Dataset 5 contains 259 variables and 961 cases (villages). Dataset 6 contains 84 variables and 37,969 cases (individual rural household members). Dataset 7 contains 449 variables and 9,200 cases (rural households). Dataset 8 contains 38 variables and 8,121 cases (individual school-age children). Dataset 9 contains 76 variables and 5,327 cases (individual rural-urban migrant household members). Dataset 10 contains 129 variables and 2,000 cases (rural-urban migrant households). The Chinese Household Income Project collected data in 1988, 1995, 2002, and 2007. ICPSR holds data from the first three collections, and information about these can be found on the series description page. Data collected in 2007 are available through the China Institute for Income Distribution. The purpose of this project was to measure and estimate the distribution of personal income in both rural and urban areas of the People's Republic of China. The study was interview-based. Five main questionnaire forms (Urban, Rural, Rural Migrant, Social Network, and Village) were filled in by interviewers at the various locations, based on questions asked of respondents. Individuals were not all interviewed directly; household members were allowed to answer questions on behalf of other members. In addition, interviewers made some direct observations about the households. Respondents in datasets 1-4 and 6-10 were members and heads of households. In dataset 5, respondents were village representatives: for each village, interviewers asked questions of the party branch secretary, the head of the village committee, or the village accountant. Village authorities were encouraged to use existing statistical data where it was available. All datasets contain a wide range of demographic and economic variables, including income, assets, liabilities, and expenditures. Cases are coded such that individuals can be linked to the information about their households and villages in other datasets. Datasets about individuals (datasets 1, 6, and 9) all include demographic variables such as household composition, gender, age, nationality, marital status, party membership, educational history, and health information. Dataset 1 is about individuals living in urban areas. It contains standard demographic variables as well as economic variables such as medical insurance and expenditures, economically productive social contacts, and employment information including occupation, sector, income, hours, conditions, job history, and training. Dataset 2 is about households in urban areas...

Study information

The sample included in this dataset represents five children who participated in a number line intervention study. Originally six children were included in the study, but one of them fulfilled the criterion for exclusion after missing several consecutive sessions. Thus, their data is not included in the dataset.

All participants were currently attending Year 1 of primary school at an independent school in New South Wales, Australia. For children to be able to eligible to participate they had to present with low mathematics achievement by performing at or below the 25th percentile in the Maths Problem Solving and/or Numerical Operations subtests from the Wechsler Individual Achievement Test III (WIAT III A & NZ, Wechsler, 2016). Participants were excluded from participating if, as reported by their parents, they have any other diagnosed disorders such as attention deficit hyperactivity disorder, autism spectrum disorder, intellectual disability, developmental language disorder, cerebral palsy or uncorrected sensory disorders.

The study followed a multiple baseline case series design, with a baseline phase, a treatment phase, and a post-treatment phase. The baseline phase varied between two and three measurement points, the treatment phase varied between four and seven measurement points, and all participants had 1 post-treatment measurement point.

The number of measurement points were distributed across participants as follows:

Participant 1 – 3 baseline, 6 treatment, 1 post-treatment

Participant 3 – 2 baseline, 7 treatment, 1 post-treatment

Participant 5 – 2 baseline, 5 treatment, 1 post-treatment

Participant 6 – 3 baseline, 4 treatment, 1 post-treatment

Participant 7 – 2 baseline, 5 treatment, 1 post-treatment

In each session across all three phases children were assessed in their performance on a number line estimation task, a single-digit computation task, a multi-digit computation task, a dot comparison task and a number comparison task. Furthermore, during the treatment phase, all children completed the intervention task after these assessments. The order of the assessment tasks varied randomly between sessions.

Measures

Number Line Estimation. Children completed a computerised bounded number line task (0-100). The number line is presented in the middle of the screen, and the target number is presented above the start point of the number line to avoid signalling the midpoint (Dackermann et al., 2018). Target numbers included two non-overlapping sets (trained and untrained) of 30 items each. Untrained items were assessed on all phases of the study. Trained items were assessed independent of the intervention during baseline and post-treatment phases, and performance on the intervention is used to index performance on the trained set during the treatment phase. Within each set, numbers were equally distributed throughout the number range, with three items within each ten (0-10, 11-20, 21-30, etc.). Target numbers were presented in random order. Participants did not receive performance-based feedback. Accuracy is indexed by percent absolute error (PAE) [(number estimated - target number)/ scale of number line] x100.

Single-Digit Computation. The task included ten additions with single-digit addends (1-9) and single-digit results (2-9). The order was counterbalanced so that half of the additions present the lowest addend first (e.g., 3 + 5) and half of the additions present the highest addend first (e.g., 6 + 3). This task also included ten subtractions with single-digit minuends (3-9), subtrahends (1-6) and differences (1-6). The items were presented horizontally on the screen accompanied by a sound and participants were required to give a verbal response. Participants did not receive performance-based feedback. Performance on this task was indexed by item-based accuracy.

Multi-digit computational estimation. The task included eight additions and eight subtractions presented with double-digit numbers and three response options. None of the response options represent the correct result. Participants were asked to select the option that was closest to the correct result. In half of the items the calculation involved two double-digit numbers, and in the other half one double and one single digit number. The distance between the correct response option and the exact result of the calculation was two for half of the trials and three for the other half. The calculation was presented vertically on the screen with the three options shown below. The calculations remained on the screen until participants responded by clicking on one of the options on the screen. Participants did not receive performance-based feedback. Performance on this task is measured by item-based accuracy.

Dot Comparison and Number Comparison. Both tasks included the same 20 items, which were presented twice, counterbalancing left and right presentation. Magnitudes to be compared were between 5 and 99, with four items for each of the following ratios: .91, .83, .77, .71, .67. Both quantities were presented horizontally side by side, and participants were instructed to press one of two keys (F or J), as quickly as possible, to indicate the largest one. Items were presented in random order and participants did not receive performance-based feedback. In the non-symbolic comparison task (dot comparison) the two sets of dots remained on the screen for a maximum of two seconds (to prevent counting). Overall area and convex hull for both sets of dots is kept constant following Guillaume et al. (2020). In the symbolic comparison task (Arabic numbers), the numbers remained on the screen until a response was given. Performance on both tasks was indexed by accuracy.

The Number Line Intervention

During the intervention sessions, participants estimated the position of 30 Arabic numbers in a 0-100 bounded number line. As a form of feedback, within each item, the participants’ estimate remained visible, and the correct position of the target number appeared on the number line. When the estimate’s PAE was lower than 2.5, a message appeared on the screen that read “Excellent job”, when PAE was between 2.5 and 5 the message read “Well done, so close! and when PAE was higher than 5 the message read “Good try!” Numbers were presented in random order.

Variables in the dataset

Age = age in ‘years, months’ at the start of the study

Sex = female/male/non-binary or third gender/prefer not to say (as reported by parents)

Math_Problem_Solving_raw = Raw score on the Math Problem Solving subtest from the WIAT III (WIAT III A & NZ, Wechsler, 2016).

Math_Problem_Solving_Percentile = Percentile equivalent on the Math Problem Solving subtest from the WIAT III (WIAT III A & NZ, Wechsler, 2016).

Num_Ops_Raw = Raw score on the Numerical Operations subtest from the WIAT III (WIAT III A & NZ, Wechsler, 2016).

Math_Problem_Solving_Percentile = Percentile equivalent on the Numerical Operations subtest from the WIAT III (WIAT III A & NZ, Wechsler, 2016).

The remaining variables refer to participants’ performance on the study tasks. Each variable name is composed by three sections. The first one refers to the phase and session. For example, Base1 refers to the first measurement point of the baseline phase, Treat1 to the first measurement point on the treatment phase, and post1 to the first measurement point on the post-treatment phase.

The second part of the variable name refers to the task, as follows:

DC = dot comparison

SDC = single-digit computation

NLE_UT = number line estimation (untrained set)

NLE_T= number line estimation (trained set)

CE = multidigit computational estimation

NC = number comparison

The final part of the variable name refers to the type of measure being used (i.e., acc = total correct responses and pae = percent absolute error).

Thus, variable Base2_NC_acc corresponds to accuracy on the number comparison task during the second measurement point of the baseline phase and Treat3_NLE_UT_pae refers to the percent absolute error on the untrained set of the number line task during the third session of the Treatment phase.

The PATH Study was launched in 2011 to inform the Food and Drug Administration's regulatory activities under the Family Smoking Prevention and Tobacco Control Act (TCA). The PATH Study is a collaboration between the National Institute on Drug Abuse (NIDA), National Institutes of Health (NIH), and the Center for Tobacco Products (CTP), Food and Drug Administration (FDA). The study sampled over 150,000 mailing addresses across the United States to create a national sample of tobacco users and non-users. 45,971 adults and youth constitute the first (baseline) wave, Wave 1, of data collected by this longitudinal cohort study. These 45,971 adults and youth along with 7,207 "shadow youth" (youth ages 9 to 11 sampled at Wave 1) make make up the 53,178 participants that constitute the Wave 1 Cohort. Respondents are asked to complete an interview at each follow-up wave. Youth who turn 18 by the current wave of data collection are considered "aged-up adults" and are invited to complete the Adult Interview. Additionally, "shadow youth" are considered "aged-up youth" upon turning 12 years old, when they are asked to complete an interview after parental consent. At Wave 4, a probability sample of 14,098 adults, youth, and shadow youth ages 10 to 11 was selected from the civilian, noninstitutionalized population at the time of Wave 4. This sample was recruited from residential addresses not selected for Wave 1 in the same sampled PSUs and segments using similar within-household sampling procedures. This "replenishment sample" was combined for estimation and analysis purposes with Wave 4 adult and youth respondents from the Wave 1 Cohort who were in the civilian, noninstitutionalized population at the time of Wave 4. This combined set of Wave 4 participants, 52,731 participants in total, forms the Wave 4 Cohort. Please refer to the Restricted-Use Files User Guide that provides further details about children designated as "shadow youth" and the formation of the Wave 1 and Wave 4 Cohorts. Dataset 0001 (DS0001) contains the data from the Master Linkage file. This file contains 42 variables and 67,276 cases. The file provides a master list of every person's unique identification number and what type of respondent they were for each wave. Dataset 1011 (DS1011) contains the data from the Wave 1 Adult Questionnaire. This data file contains 2,021 variables and 32,320 cases. Each of the cases represents a single, completed interview. Dataset 1012 (DS1012) contains the data from the Wave 1 Youth (and Parent) Questionnaire. This file contains 1,431 variables and 13,651 cases. Dataset 1411 (DS1411) contains the Wave 1 State Identifier data for Adults and has 5 variables and 32,320 cases. Dataset 1412 (DS1412) contains the Wave 1 State Identifier data for Youth (and Parents) and has 5 variables and 13,651 cases. The same 5 variables are in each State Identifier dataset, including PERSONID for linking the State Identifier to the questionnaire and biomarker data and 3 variables designating the state (state FIPS, state abbreviation, and full name of the state). Dataset 2011 (DS2011) contains the data from the Wave 2 Adult Questionnaire. This data file contains 2,421 variables and 28,362 cases. Of these cases, 26,447 also completed a Wave 1 Adult Questionnaire. The other 1,915 cases are "aged-up adults" having previously completed a Wave 1 Youth Questionnaire.Dataset 2012 (DS2012) contains the data from the Wave 2 Youth (and Parent) Questionnaire. This data file contains 1,596 variables and 12,172 cases. Of these cases, 10,081 also completed a Wave 1 Youth Questionnaire. The other 2,091 cases are "aged-up youth" having previously been sampled as "shadow youth." Dataset 3011 (DS3011) contains the data from the Wave 3 Adult Questionnaire. This data file contains 2,359 variables and 28,148 cases. Of these cases, 26,241 are continuing adults having completed a prior Adult Questionnaire. The other 1,907 cases are "aged-up adults" having previously completed a Youth Questionnaire. Dataset 3012 (DS3012) contains the data from the Wave 3 Youth (and Parent) Questionnaire. This data file contains 1,492 variables and 11,814 cases. Of these cases, 9,769 are continuing youth having completed a prior Youth Interview. The other 2,045 cases are "aged-up youth" having previously been sampled as "shadow youth." Datasets 3111, 3211, 3112, and 3212 (DS3111, DS3211, DS3112, and DS3212) are data files comprising the weight variables for Wave 3. The weight variables for Wave 1 and Wave 2 are included in the main data files. However, starting with Wave 3, the weight variables have been separated into individual data files. The "all-waves" weight files contain weights for respondents who completed an interview for all waves in which they were old enough to do so or verified their information with the study for waves in which they were not old enough to be interviewed. The "single-wave" weight files contain weights for all respondents in Wave 3 regardless of their

This project was undertaken to establish a computerized skeletal database composed of recent forensic cases to represent the present ethnic diversity and demographic structure of the United States population. The intent was to accumulate a forensic skeletal sample large and diverse enough to reflect different socioeconomic groups of the general population from different geographical regions of the country in order to enable researchers to revise the standards being used for forensic skeletal identification. The database is composed of eight data files, comprising four categories. The primary "biographical" or "identification" files (Part 1, Demographic Data, and Part 2, Geographic and Death Data) comprise the first category of information and pertain to the positive identification of each of the 1,514 data records in the database. Information in Part 1 includes sex, ethnic group affiliation, birth date, age at death, height (living and cadaver), and weight (living and cadaver). Variables in Part 2 pertain to the nature of the remains, means and sources of identification, city and state/country born, occupation, date missing/last seen, date of discovery, date of death, time since death, cause of death, manner of death, deposit/exposure of body, area found, city, county, and state/country found, handedness, and blood type. The Medical History File (Part 3) represents the second category of information and contains data on the documented medical history of the individual. Variables in Part 3 include general comments on medical history as well as comments on congenital malformations, dental notes, bone lesions, perimortem trauma, and other comments. The third category consists of an inventory file (Part 4, Skeletal Inventory Data) in which data pertaining to the specific contents of the database are maintained. This includes the inventory of skeletal material by element and side (left and right), indicating the condition of the bone as either partial or complete. The variables in Part 4 provide a skeletal inventory of the cranium, mandible, dentition, and postcranium elements and identify the element as complete, fragmentary, or absent. If absent, four categories record why it is missing. The last part of the database is composed of three skeletal data files, covering quantitative observations of age-related changes in the skeleton (Part 5), cranial measurements (Part 6), and postcranial measurements (Part 7). Variables in Part 5 provide assessments of epiphyseal closure and cranial suture closure (left and right), rib end changes (left and right), Todd Pubic Symphysis, Suchey-Brooks Pubic Symphysis, McKern & Steward--Phases I, II, and III, Gilbert & McKern--Phases I, II, and III, auricular surface, and dorsal pubic pitting (all for left and right). Variables in Part 6 include cranial measurements (length, breadth, height) and mandibular measurements (height, thickness, diameter, breadth, length, and angle) of various skeletal elements. Part 7 provides postcranial measurements (length, diameter, breadth, circumference, and left and right, where appropriate) of the clavicle, scapula, humerus, radius, ulna, scarum, innominate, femur, tibia, fibula, and calcaneus. A small file of noted problems for a few cases is also included (Part 8).

Data Set Information

This data set contains records related to red and white variants of the Portuguese Vinho Verde wine. It contains information from 1599 red wine samples and 4898 white wine samples. Input variables in the data set consist of the type of wine (either red or white wine) and metrics from objective tests (e.g. acidity levels, PH values, ABV, etc.), while the target/output variable is a numerical score based on sensory data—median of at least 3 evaluations made by wine experts. Each expert graded the wine quality between 0 (very bad) and 10 (very excellent). Due to privacy and logistic issues, there is no data about grape types, wine brand, and wine selling price.

This data set is a combined version of the two separate files (distinct red and white wine data sets) originally shared in the UCI Machine Learning Repository.

The following are some existing data sets on Kaggle from the same source (with notable differences from this data set): - Red Wine Quality (contains red wine data only) - Wine Quality (combination of red and white wine data but with some values randomly removed) - Wine Quality (red and white wine data not combined)

Contents

Input variables:

1 - type of wine: type of wine (categorical: 'red', 'white')

(continuous variables based on physicochemical tests)

2 - fixed acidity: The acids that naturally occur in the grapes used to ferment the wine and carry over into the wine. They mostly consist of tartaric, malic, citric or succinic acid that mostly originate from the grapes used to ferment the wine. They also do not evaporate easily. (g / dm^3)

3 - volatile acidity: Acids that evaporate at low temperatures—mainly acetic acid which can lead to an unpleasant, vinegar-like taste at very high levels. (g / dm^3)

4 - citric acid: Citric acid is used as an acid supplement which boosts the acidity of the wine. It's typically found in small quantities and can add 'freshness' and flavor to wines. (g / dm^3)

5 - residual sugar: The amount of sugar remaining after fermentation stops. It's rare to find wines with less than 1 gram/liter. Wines residual sugar level greater than 45 grams/liter are considered sweet. On the other end of the spectrum, a wine that does not taste sweet is considered as dry. (g / dm^3)

6 - chlorides: The amount of chloride salts (sodium chloride) present in the wine. (g / dm^3)

7 - free sulfur dioxide: The free form of SO2 exists in equilibrium between molecular SO2 (as a dissolved gas) and bisulfite ion; it prevents microbial growth and the oxidation of wine. All else constant, the higher the free sulfur dioxide content, the stronger the preservative effect. (mg / dm^3)

8 - total sulfur dioxide: The amount of free and bound forms of S02; in low concentrations, SO2 is mostly undetectable in wine, but at free SO2 concentrations over 50 ppm, SO2 becomes evident in the nose and taste of wine. (mg / dm^3)

9 - density: The density of wine juice depending on the percent alcohol and sugar content; it's typically similar but higher than that of water (wine is 'thicker'). (g / cm^3)

10 - pH: A measure of the acidity of wine; most wines are between 3-4 on the pH scale. The lower the pH, the more acidic the wine is; the higher the pH, the less acidic the wine. (The pH scale technically is a logarithmic scale that measures the concentration of free hydrogen ions floating around in your wine. Each point of the pH scale is a factor of 10. This means a wine with a pH of 3 is 10 times more acidic than a wine with a pH of 4)

11 - sulphates: Amount of potassium sulphate as a wine additive which can contribute to sulfur dioxide gas (S02) levels; it acts as an antimicrobial and antioxidant agent.(g / dm3)

12 - alcohol: How much alcohol is contained in a given volume of wine (ABV). Wine generally contains between 5–15% of alcohols. (% by volume)

Output variable:

13 - quality: score between 0 (very bad) and 10 (very excellent) by wine experts

Acknowledgements

Source: P. Cortez, A. Cerdeira, F. Almeida, T. Matos and J. Reis. Modeling wine preferences by data mining from physicochemical properties. In Decision Support Systems, Elsevier, 47(4):547-553, 2009.

Data credit goes to UCI. Visit their website to access the original data set directly: https://archive.ics.uci.edu/ml/datasets/wine+quality

Context

So much about wine making remains elusive—taste is very subjective, making it extremely challenging to predict exactly how consumers will react to a certain bottle of wine. There is no doubt that winemakers, connoisseurs, and scientists have greatly contributed their expertise to ...

This zip file contains: - 3 .zip files = projects to be imported into SmartPLS 3

DLOQ-A model with 7 dimensions DLOQ-A model with second-order latent variable ECSI model (Tenenhaus et al., 2005) to exemplify direct, indirect and total effects, as well as importance-performance map and moderation with continuous variables. ECSI Model (Sanches, 2013) to exemplify MGA (multi-group analysis)

• 5 files (csv, txt) with data to run 7 examples in SmartPLS 3

Note: - DLOQ-A = new dataset (ours) - ECSI-Tenenhaus et al. [model for mediation and moderation] = available at: http://www.smartpls.com > Resources > SmartPLS Project Examples - ECSI-Sanches [dataset for MGA] = available in the software R > library(plspm) > data(satisfaction)

UCI-Bank-Marketing-Dataset

The following information is drawn from the UCI Machine Learning Repository: https://archive.ics.uci.edu/ml/datasets/bank+marketing

Abstract

The data is related with direct marketing campaigns (phone calls) of a Portuguese banking institution. The classification goal is to predict if the client will subscribe a term deposit (variable y).

Source

[Moro et al., 2014] S. Moro, P. Cortez and P. Rita. A Data-Driven Approach to Predict the Success of Bank Telemarketing. Decision Support Systems, Elsevier, 62:22-31, June 2014

Data Set Information

The data is related with direct marketing campaigns of a Portuguese banking institution. The marketing campaigns were based on phone calls. Often, more than one contact to the same client was required, in order to access if the product (bank term deposit) would be ('yes') or not ('no') subscribed.

There are four datasets:

1) bank-additional-full.csv with all examples (41188) and 20 inputs, ordered by date (from May 2008 to November 2010), very close to the data analyzed in [Moro et al., 2014] <-- this is used 2) bank-additional.csv with 10% of the examples (4119), randomly selected from 1), and 20 inputs. 3) bank-full.csv with all examples and 17 inputs, ordered by date (older version of this dataset with less inputs). 4) bank.csv with 10% of the examples and 17 inputs, randomly selected from 3 (older version of this dataset with less inputs).

The classification goal is to predict if the client will subscribe (yes/no) a term deposit (variable y).

Attribute information

Input variables

Bank client data

1. age (numeric)
2. job : type of job (categorical)
3. marital : marital status (categorical)
4. education (categorical)
5. default: has credit in default? (categorical)
6. housing: has housing loan? (categorical)
7. loan: has personal loan? (categorical)

Related with the last contact of the current campaign

1. contact: contact communication type (categorical)
2. month: last contact month of year (categorical)
3. day_of_week: last contact day of the week (categorical)
4. duration: last contact duration, in seconds (numeric). Important note: this attribute highly affects the output target (e.g., if duration=0 then y='no'). Yet, the duration is not known before a call is performed. Also, after the end of the call y is obviously known. Thus, this input should only be included for benchmark purposes and should be discarded if the intention is to have a realistic predictive model.

Other attributes

1. campaign: number of contacts performed during this campaign and for this client (numeric, includes last contact)
2. pdays: number of days that passed by after the client was last contacted from a previous campaign (numeric; 999 means client was not previously contacted)
3. previous: number of contacts performed before this campaign and for this client (numeric)
4. poutcome: outcome of the previous marketing campaign (categorical)

Social and economic context attributes

1. emp.var.rate: employment variation rate - quarterly indicator (numeric)
2. cons.price.idx: consumer price index - monthly indicator (numeric)
3. cons.conf.idx: consumer confidence index - monthly indicator (numeric)
4. euribor3m: euribor 3 month rate - daily indicator (numeric)
5. nr.employed: number of employees - quarterly indicator (numeric)

Output variable (desired target)

1. y - has the client subscribed a term deposit? (binary: 'yes','no')

Relevant Papers

S. Moro, P. Cortez and P. Rita. A Data-Driven Approach to Predict the Success of Bank Telemarketing. Decision Support Systems, Elsevier, 62:22-31, June 2014

S. Moro, R. Laureano and P. Cortez. Using Data Mining for Bank Direct Marketing: An Application of the CRISP-DM Methodology. In P. Novais et al. (Eds.), Proceedings of the European Simulation and Modelling Conference - ESM'2011, pp. 117-121, Guimaraes, Portugal, October, 2011. EUROSIS. [bank.zip]

Banner image

Photo by Karolina Grabowska from Pexels

This repository provides the preprocessed datasets, which are used in the study Temperature forecasting by deep learning methods by Gong et al. (2022). This allows the user to reproduce the presented results without running the preprocessing chain from the raw ERA5 data. Data description The datasets used to train, validate, and test the deep neural networks are based on the ERA5 reanalysis data provided by the European Centre for Medium-range Weather Forecast (ECMWF). Five different datasets have been created. All incorporate data between the years 2007 and 2019, but cover slightly varying domains over Central Europe and include different meteorological variables. The datasets are made available in compressed tar-archives (see Storage Location URL below). The file names thereby encapsulate some meta-information using the following naming convention: ERA5-Y[yyyy]-[yyyy]M[mm]to[mm]-[nx]x[ny]-[nn.nn]N[ee.ee]E-[var1]_[var2]_[var3] where - Y[yyyy]-[yyyy]M[mm]to[mm] denotes the years and the months describing the data period, -[nx]x[ny] is the number of grid points/pixels of the target domain in longitude and latitude direction, -[nn.nn]N[ee.ee]E stands for the geographical coordinates in degree of the target domain's south-west corner and -[var1]_[var2]_[var3] denote the short names of the variables according to ECMWF's parameter database In particular, the following datasets are provided: 1) era5-Y2007-2019M01to12-92x56-3840N0000E-2t_tcc_t850.tar.bz2: The target domain extends from 38.4°N to 54.9°N and 0.0°E to 27.3°E (92x56 grid points). The 2m-temperature (2t), the total cloud cover (tcc), and the 850 hPa temperature (t_850) are included as variables. This data corresponds to Datasets ID 1-3 in table A1 of the manuscript. 2) era5-Y2007-2019M01to12-80x48-3960N0180E-2t_tcc_t850.tar.bz2: The target domain extends from 39.6°N to 53.7°N and 1.8°E to 25.5°E (80x48 grid points). The 2t, tcc, and the t_850 are included as variables. This data corresponds to Dataset ID 4 in table A1 of the manuscript. 3) era5-Y2007-2019M01to12-72x44-4020N0300E-2t_tcc_t_850.tar.bz2: The target domain extends from 40.2°N to 53.1°N and 3.0°E to 24.3°E (72x44 grid points). The 2t, tcc, and t_850 are included as variables. This data corresponds to Dataset ID 5 in table A1 of the manuscript. 4) era5-Y2007-2019M01to12-80x48-3960N0180E-2t_t850.tar.bz2: The target domain extends from 39.6°N to 53.7°N and 1.8°E to 25.5°E (80x48 grid points). The 2t and the t_850 are the only variables included. This data set is actually a subset of No. 2. This data corresponds to Dataset ID 6 in table A1 of the manuscript. 5) era5-Y2007-2019M01to12-80x48-3960N0180E-2t.tar.bz2: The target domain extends from 39.6°N to 53.7°N and 1.8°E to 25.5°E (80x48 grid points). 2t is exclusively included. This data set is also a subset of No. 2. This data corresponds to Dataset ID 7 in table A1 of the manuscript. Data creation The original ERA5 data can be retrieved from the (MARS archive). Once access is granted, data can be downloaded by specifying a resolution of 0.3° in the retrieval script. The datasets provided in this repository are the processed ERA5 data after the extraction and the two preprocessing steps using the Atmospheric Machine learning Benchmarking System (AMBS) workflow tool (more details are provided in the README of the corresponding code repository). The data is available in TFRecords format that is used directly in the training step. Data access and decompression Data are stored in the archived and compressed format tar.bz2 and available via: https://datapub.fz-juelich.de/esde/esde-nfs/online_publication/2mT_by_DL/. After downloading, the compressed archives can be unpacked on Linux using tar xjf [filename].tar.bz2. On Windows, decompressing can be performed using WinZip. Dataset content After decompressing, the following subdirectory structure is created from each compressed tar-archive: - tfrecords_seq_len_[sequence_length]: This folder holds the TFRecords files that are streamed to the deep neural networks during training and postprocessing. Each TFRecord file contains 10 samples, where each sample comprises a sequence over [sequence_length] hours. - pickle: This folder contains the normalized hourly data saved in monthly pickle files (X_[month].pkl). The corresponding timestamps are included in T_[month].pkl. Furthermore, statistical information for each month is provided in the files stat_[month].json. - metadata.json: This file provides important meta information including the coordinates of the target domain, the included variables (e.g. 2t and t_850) and the origin of the processed data. - statsitic.json: This file includes the statistical information (maximum, minimum, and average values) used for normalizing the data. It also includes other information such as the total number of the timestamps (nfiles) and the list of JSON files (stat_[month].json) to compute the statistics. Data integrity and verification The tar-archives have been recursively checksummed with the md5 hash function. The generated file is uploaded to ensure the integrity of the files and no alteration to the dataset. To verify the integrity of the downloaded data, use the following snippet: find -type f -exec md5sum '{}' \; > md5sum.txt It will generate a single text file that should be identical to the file in this entry. License Original data by ECMWF Copyright "© 2022 European Centre for Medium-Range Weather Forecasts (ECMWF)". Source www.ecmwf.int. This data is published under the Creative Commons Attribution 4.0 International (CC BY 4.0) license. https://creativecommons.org/licenses/by/4.0/ Contact Bing Gong (b.gong@fz-juelich.de)

Source:

[Moro et al., 2014] S. Moro, P. Cortez and P. Rita. A Data-Driven Approach to Predict the Success of Bank Telemarketing. Decision Support Systems, Elsevier, 62:22-31, June 2014

Data Set Information:

The data is related with direct marketing campaigns of a Portuguese banking institution. The marketing campaigns were based on phone calls. Often, more than one contact to the same client was required, in order to access if the product (bank term deposit) would be ('yes') or not ('no') subscribed.

There are four datasets: 1) bank-additional-full.csv with all examples (41188) and 20 inputs, ordered by date (from May 2008 to November 2010), very close to the data analyzed in [Moro et al., 2014] 2) bank-additional.csv with 10% of the examples (4119), randomly selected from 1), and 20 inputs. 3) bank-full.csv with all examples and 17 inputs, ordered by date (older version of this dataset with less inputs). 4) bank.csv with 10% of the examples and 17 inputs, randomly selected from 3 (older version of this dataset with less inputs). The smallest datasets are provided to test more computationally demanding machine learning algorithms (e.g., SVM).

The classification goal is to predict if the client will subscribe (yes/no) a term deposit (variable y).

Attribute Information:

Input variables:

bank client data:

1 - age (numeric) 2 - job : type of job (categorical: 'admin.','blue-collar','entrepreneur','housemaid','management','retired','self-employed','services','student','technician','unemployed','unknown') 3 - marital : marital status (categorical: 'divorced','married','single','unknown'; note: 'divorced' means divorced or widowed) 4 - education (categorical: 'basic.4y','basic.6y','basic.9y','high.school','illiterate','professional.course','university.degree','unknown') 5 - default: has credit in default? (categorical: 'no','yes','unknown') 6 - housing: has housing loan? (categorical: 'no','yes','unknown') 7 - loan: has personal loan? (categorical: 'no','yes','unknown')

related with the last contact of the current campaign:

8 - contact: contact communication type (categorical: 'cellular','telephone') 9 - month: last contact month of year (categorical: 'jan', 'feb', 'mar', ..., 'nov', 'dec') 10 - day_of_week: last contact day of the week (categorical: 'mon','tue','wed','thu','fri') 11 - duration: last contact duration, in seconds (numeric). Important note: this attribute highly affects the output target (e.g., if duration=0 then y='no'). Yet, the duration is not known before a call is performed. Also, after the end of the call y is obviously known. Thus, this input should only be included for benchmark purposes and should be discarded if the intention is to have a realistic predictive model.

other attributes:

12 - campaign: number of contacts performed during this campaign and for this client (numeric, includes last contact) 13 - pdays: number of days that passed by after the client was last contacted from a previous campaign (numeric; 999 means client was not previously contacted) 14 - previous: number of contacts performed before this campaign and for this client (numeric) 15 - poutcome: outcome of the previous marketing campaign (categorical: 'failure','nonexistent','success')

social and economic context attributes

16 - emp.var.rate: employment variation rate - quarterly indicator (numeric) 17 - cons.price.idx: consumer price index - monthly indicator (numeric) 18 - cons.conf.idx: consumer confidence index - monthly indicator (numeric) 19 - euribor3m: euribor 3 month rate - daily indicator (numeric) 20 - nr.employed: number of employees - quarterly indicator (numeric)

Output variable (desired target): 21 - y - has the client subscribed a term deposit? (binary: 'yes','no')

Relevant Papers:

S. Moro, P. Cortez and P. Rita. A Data-Driven Approach to Predict the Success of Bank Telemarketing. Decision Support Systems, Elsevier, 62:22-31, June 2014

S. Moro, R. Laureano and P. Cortez. Using Data Mining for Bank Direct Marketing: An Application of the CRISP-DM Methodology. In P. Novais et al. (Eds.), Proceedings of the European Simulation and Modelling Conference - ESM'2011, pp. 117-121, Guimaraes, Portugal, October, 2011. EUROSIS. [bank.zip]

Citation Request:

This dataset is public available for research. The details are described in [Moro et al., 2014]. Please include this citation if you plan to use this database:

[Moro et al., 2014] S. Moro, P. Cortez and P. Rita. A Data-Driven Approach to Predict the Success of Bank Telemarketing. Decision Support Systems, Elsevier, 62:22-31, June 2014