#https://www.kaggle.com/c/facial-keypoints-detection/details/getting-started-with-r #################################

###Variables for downloaded files data.dir <- ' ' train.file <- paste0(data.dir, 'training.csv') test.file <- paste0(data.dir, 'test.csv') #################################

###Load csv -- creates a data.frame matrix where each column can have a different type. d.train <- read.csv(train.file, stringsAsFactors = F) d.test <- read.csv(test.file, stringsAsFactors = F)

###In training.csv, we have 7049 rows, each one with 31 columns. ###The first 30 columns are keypoint locations, which R correctly identified as numbers. ###The last one is a string representation of the image, identified as a string.

###To look at samples of the data, uncomment this line:

head(d.train)

###Let's save the first column as another variable, and remove it from d.train: ###d.train is our dataframe, and we want the column called Image. ###Assigning NULL to a column removes it from the dataframe

im.train <- d.train$Image d.train$Image <- NULL #removes 'image' from the dataframe

im.test <- d.test$Image d.test$Image <- NULL #removes 'image' from the dataframe

################################# #The image is represented as a series of numbers, stored as a string #Convert these strings to integers by splitting them and converting the result to integer

#strsplit splits the string #unlist simplifies its output to a vector of strings #as.integer converts it to a vector of integers. as.integer(unlist(strsplit(im.train[1], " "))) as.integer(unlist(strsplit(im.test[1], " ")))

###Install and activate appropriate libraries ###The tutorial is meant for Linux and OSx, where they use a different library, so: ###Replace all instances of %dopar% with %do%.

install.packages('foreach')

library("foreach", lib.loc="~/R/win-library/3.3")

###implement parallelization im.train <- foreach(im = im.train, .combine=rbind) %do% { as.integer(unlist(strsplit(im, " "))) } im.test <- foreach(im = im.test, .combine=rbind) %do% { as.integer(unlist(strsplit(im, " "))) } #The foreach loop will evaluate the inner command for each row in im.train, and combine the results with rbind (combine by rows). #%do% instructs R to do all evaluations in parallel. #im.train is now a matrix with 7049 rows (one for each image) and 9216 columns (one for each pixel):

###Save all four variables in data.Rd file ###Can reload them at anytime with load('data.Rd')

save(d.train, im.train, d.test, im.test, file='data.Rd')

load('data.Rd')

#each image is a vector of 96*96 pixels (96*96 = 9216). #convert these 9216 integers into a 96x96 matrix: im <- matrix(data=rev(im.train[1,]), nrow=96, ncol=96)

#im.train[1,] returns the first row of im.train, which corresponds to the first training image. #rev reverse the resulting vector to match the interpretation of R's image function #(which expects the origin to be in the lower left corner).

#To visualize the image we use R's image function: image(1:96, 1:96, im, col=gray((0:255)/255))

#Let’s color the coordinates for the eyes and nose points(96-d.train$nose_tip_x[1], 96-d.train$nose_tip_y[1], col="red") points(96-d.train$left_eye_center_x[1], 96-d.train$left_eye_center_y[1], col="blue") points(96-d.train$right_eye_center_x[1], 96-d.train$right_eye_center_y[1], col="green")

#Another good check is to see how variable is our data. #For example, where are the centers of each nose in the 7049 images? (this takes a while to run): for(i in 1:nrow(d.train)) { points(96-d.train$nose_tip_x[i], 96-d.train$nose_tip_y[i], col="red") }

#there are quite a few outliers -- they could be labeling errors. Looking at one extreme example we get this: #In this case there's no labeling error, but this shows that not all faces are centralized idx <- which.max(d.train$nose_tip_x) im <- matrix(data=rev(im.train[idx,]), nrow=96, ncol=96) image(1:96, 1:96, im, col=gray((0:255)/255)) points(96-d.train$nose_tip_x[idx], 96-d.train$nose_tip_y[idx], col="red")

#One of the simplest things to try is to compute the mean of the coordinates of each keypoint in the training set and use that as a prediction for all images colMeans(d.train, na.rm=T)

#To build a submission file we need to apply these computed coordinates to the test instances: p <- matrix(data=colMeans(d.train, na.rm=T), nrow=nrow(d.test), ncol=ncol(d.train), byrow=T) colnames(p) <- names(d.train) predictions <- data.frame(ImageId = 1:nrow(d.test), p) head(predictions)

#The expected submission format has one one keypoint per row, but we can easily get that with the help of the reshape2 library:

install.packages('reshape2')

library(...

If these look unintelligible to you, that’s okay! They are only meant to show that in R, there are different syntax strategies to complete the same tasks. If one looks more interpretable than the others, then great—you have found your coding style to begin!

This R code uses joint time series of concentration and discharge to (1) separate discharge events and store them in a data frame (events_h) and (2) analyse C-Q relationships including hysteresis, derive metrics describing these and store them in a data frame (eve.des). The R code is provided as TXT and as R-file. The code is written by Qing Zhan, Rémi Dupas, Camille Minaudo and Andreas Musolff.

This code is used and further descriped in this paper: A. Musolff, Q. Zhan, R. Dupas, C. Minaudo, J. H. Fleckenstein, M. Rode, J. Dehaspe & K. Rinke (2021) Spatial and Temporal Variability in Concentration-Discharge Relationships at the Event Scale. Water Resourcers Research Volume 57, Issue 10 https://doi.org/10.1029/2020WR029442

• Main objective : To forecast the page visits of a website
• Tool : Time Series Forecasting using Prophet in R.
• Steps:
• Read the data
• Data Cleaning: Checking data types, date formats and missing data https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F10868729%2F56d7b1edf4f51157804e81b02c032e4d%2FPicture1.png?generation=1690271521103777&alt=media" alt="">
• Run libraries (dplyr, ggplot2, tidyverse, lubridate, prophet, forecast)
• Change the Date column from character vector to date and change data format using lubridate package
• Rename the column "Date" to "ds" and "Visits" to "y".
• Treat "Christmas" and "Black.Friday" as holiday events. As the data ranges from 2016 to 2020, there will be 5 Christmas and 5 Black Friday days.
• We will look at the impact of Christmas 3 days prior and 3 days later from Christmas date on "Visits" and 3 days prior and 1 day later for Black Friday
• We create two data frames called Christmas and Black.Friday and merge the two into a data frame called "holidays". https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F10868729%2Fd07b366be2050fefe6a62563b6abac0c%2FPicture2.png?generation=1690272066356516&alt=media" alt="">
• We create train and test data. In train data & test data, we select only 3 variables namely ds, y , Easter. In train data, ds contains data before 2020-12-01 and test data contains data equal to and after 2020-12-01 (31 days) data
• Train Data
• https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F10868729%2F8f3f58fe40b29b276bb7103cb1dfdde1%2FPicture3.png?generation=1690272272038405&alt=media" alt="">
• Test Data
• https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F10868729%2Fb4362117f46aeb210dad23f07d3ecb39%2FPicture4.png?generation=1690272400355824&alt=media" alt="">
• Use prophet model which will include multiple parameter. We are going with the default parameters. Thereafter, we add the external regressor "Easter".
• https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F10868729%2F7325be63d887372cc5764ddf29a94310%2FPicture5.png?generation=1690272892963939&alt=media" alt="">
• We create the future data frame for forecasting and name the data frame "future". It will include "m" and 31 days of the test data. We then predict this future data frame and create a new data frame called "forecast".
• Forecast data frame consists of 1827 rows and 34 variables. This shows the external Regressor (Easter) value is 0 through the entire time period. This shows that "Easter" has no impact or effect on "Visits".
• yhat stands for the predicted value (predicted visits).
• https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F10868729%2Fae5c9414d1b1bbb2670b372a326970a5%2FPicture6.png?generation=1690273558489681&alt=media" alt="">
• We try to understand the impact of Holiday events "Christmas" and "Black.Friday"
• https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F10868729%2F5a36cc5308f9e46f0b63fa8e37c4b932%2FPicture7.png?generation=1690273814760538&alt=media" alt="">
• https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F10868729%2F8cc3dd0581db1e8b9d542d9a524abd39%2FPicture8.png?generation=1690273879506571&alt=media" alt="">
• We plot the forecast.
• plot(m,forecast) https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F10868729%2Fa7968ff05abdd5b4e789f3723b41c4ed%2FPicture9.png?generation=1690274020880594&alt=media" alt="">
• blue is predicted value(yhat) and black is actual value(y) and blue shaded regions are the yhat_upper and yhat_lower values
• prophet_plot_components(m,forecast) https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F10868729%2F52408afb8c71118ef6729420085875e8%2FPicture10.png?generation=1690274184325240&alt=media" alt="">
• Trend indicates that the page visits remained constant from Jan'16 to Mid'17 and thereafter there was an upswing from Mid'19 to End of 2020
• From Holidays, we can make out that Christmas had a negative effect on page visits whereas Black Friday had a positive effect on page visits
• Weekly seasonality indicates that page visits tend to remain the highest from Monday to Thursday and starts going down thereafter
• Yearly seasonality indicates that page visits are the highest in Apr and then starts going down thereafter with
• Oct having reaching the bottom point
• External regressor "Easter" has no impact on page visits
• plot(m,forecast) + add_changepoints_to_plot(m)
• https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F10868729%2F1253a0e381ae04d3156a4b098dafb2ca%2FPicture11.png?generation=1690274373570449&alt=media" alt="">
• Trend which is indicated by the red line starts moving upwards from Mid 2019 to 2020 onwards
• We check for acc...

Collection of code and dataframes used to calculate ice support vector-based metrics and run ice support models in R for Ecology submission.

Dataframes:

IceSupportDF_envinddatana.RData - Data frame used to run both distance and support models.

Code:

Final_Distance_Model_Ecology.R - Code used to run and evaluate distance model

Final_Support_Model_Ecology.R - Code used to run and evaluate support model

ice_vector_metrics_Ecology.R - Code showing mathematical functions listed in manuscript for vector based ice support metrics were calculated

User Agreement, Public Domain Dedication, and Disclaimer of Liability. By accessing or downloading the data or work provided here, you, the User, agree that you have read this agreement in full and agree to its terms. The person who owns, created, or contributed a work to the data or work provided here dedicated the work to the public domain and has waived his or her rights to the work worldwide under copyright law. You can copy, modify, distribute, and perform the work, for any lawful purpose, without asking permission. In no way are the patent or trademark rights of any person affected by this agreement, nor are the rights that any other person may have in the work or in how the work is used, such as publicity or privacy rights. Pacific Science & Engineering Group, Inc., its agents and assigns, make no warranties about the work and disclaim all liability for all uses of the work, to the fullest extent permitted by law. When you use or cite the work, you shall not imply endorsement by Pacific Science & Engineering Group, Inc., its agents or assigns, or by another author or affirmer of the work. This Agreement may be amended, and the use of the data or work shall be governed by the terms of the Agreement at the time that you access or download the data or work from this Website. Description This dataverse contains the data referenced in Rieth et al. (2017). Issues and Advances in Anomaly Detection Evaluation for Joint Human-Automated Systems. To be presented at Applied Human Factors and Ergonomics 2017. Each .RData file is an external representation of an R dataframe that can be read into an R environment with the 'load' function. The variables loaded are named ‘fault_free_training’, ‘fault_free_testing’, ‘faulty_testing’, and ‘faulty_training’, corresponding to the RData files. Each dataframe contains 55 columns: Column 1 ('faultNumber') ranges from 1 to 20 in the “Faulty” datasets and represents the fault type in the TEP. The “FaultFree” datasets only contain fault 0 (i.e. normal operating conditions). Column 2 ('simulationRun') ranges from 1 to 500 and represents a different random number generator state from which a full TEP dataset was generated (Note: the actual seeds used to generate training and testing datasets were non-overlapping). Column 3 ('sample') ranges either from 1 to 500 (“Training” datasets) or 1 to 960 (“Testing” datasets). The TEP variables (columns 4 to 55) were sampled every 3 minutes for a total duration of 25 hours and 48 hours respectively. Note that the faults were introduced 1 and 8 hours into the Faulty Training and Faulty Testing datasets, respectively. Columns 4 to 55 contain the process variables; the column names retain the original variable names. Acknowledgments. This work was sponsored by the Office of Naval Research, Human & Bioengineered Systems (ONR 341), program officer Dr. Jeffrey G. Morrison under contract N00014-15-C-5003. The views expressed are those of the authors and do not reflect the official policy or position of the Office of Naval Research, Department of Defense, or US Government.

Datasets and R scripts for the analysis in Ogino M, Strauss E, Farine D. 2022. Challenges of mismatching timescales in longitudinal studies of collective behaviour (doi.org/10.1098/rstb.2022.0064). 20201001_20201031, 20201101_20201130, 10101201_20211231, 20210101_20210131, 20210201_20210228, 20210301_20210331: files containing daily averaged GPS pairwise distance between males for each month Seasons_Monthly.csv: dataset containing the metadata for each sampling periods Census.Rdata: dataset (list) with dataframes for individual metadata (data$Birds) and census observation data (data$ind_obs, data$grp_obs). The grp_obs dataframe contains the metadata for the ind_obs dataset. Script1_GroupDetection.r: code to detect group memberships over time, using different approaches Script2_GPSpairwisedistance.R: code to produce the boxplot showing averaged GPS pairwise distances within detected communities and GLM to quantify how methodological processes applied in different approaches drive differences in cohesiveness of detected groups. This code requires the outputs of Script1. Script3_GroupSize.R: code to produce the boxplots showing the distribution of detected social unit sizes and Jaccard similarity between social units in consecutive sampling periods, and GLM to quantidy how methodological processes drive differences in group size and Jaccard similarity. This code requires the outputs of Script1 and Script2.

The CO2 dataframe is a dataset built into R showing the results of an experiment on the cold tolerance of grass. Grass samples from two regions were grown in either a chilled or nonchilled environment, and their CO2 uptake rate was tested.The dataset has been downloaded as a .csv file.

The two types of region chosen for this experiment are Quebec and Mississippi. Each type has three different plants used for this experiment. Each plant has been treated in either a chilled and nonchilled environment. Average concentration is found to be constant for all categories. Thus, plots are made with respect to variation in average CO2 uptake.

This is the studentlife dataset, converted from it's original form (hosted at https://studentlife.cs.dartmouth.edu/dataset.html) into a series of R tibbles (which are similar to a data.frame) and stored in the RData format, for compression / speed and ease of use with R. Note, in this RData format the dataset takes up much less space than the original.

These tibbles are suitable for use with the studentlife R package https://github.com/frycast/studentlife which is also available on CRAN, but we recommend installing the latest version from GitHub.

Studentlife dataset reference:

Wang, Rui, Fanglin Chen, Zhenyu Chen, Tianxing Li, Gabriella Harari, Stefanie Tignor, Xia Zhou, Dror Ben-Zeev, and Andrew T. Campbell. "StudentLife: Assessing Mental Health, Academic Performance and Behavioral Trends of College Students using Smartphones." In Proceedings of the ACM Conference on Ubiquitous Computing. 2014.

Overview This dataset contains biologging data and R script used to produce the results in "A summer heatwave reduced activity, heart rate and autumn body mass in a cold-adapted ungulate", a submitted manuscript. The longitudinal data of female reindeer and calf body masses used in the paper is owned by the Finnish Reindeer Herders’ Association. Natural Resources Institute Finland (Luke) updates, saves and administrates this long-term reindeer herd data. Methods of data collection Animals and study area The study involved biologging (see below) 14 adult semi-domesticated reindeer females (Focal animals: Table S1) at the Kutuharju Reindeer Research Facility (Kaamanen, Northern Finland, 69° 8’ N, 26° 59’ E, Figure S1), during June–September 2018. Ten of these individuals had been intensively handled in June as part of another study (Trondrud, 2021). The 14 females were part of a herd of ~100 animals, belonging to the Reindeer Herders’ Association. The herding management included keeping reindeer in two large enclosures (~13.8 and ~15 km2) after calving until the rut, after which animals were moved to a winter enclosure (~15 km2) and then in spring to a calving paddock (~0.3 km2) to give birth (See Supporting Information for further details on the study area). Kutuharju reindeer graze freely on natural pastures from May to November and after that are provided with silage and pellets as a supplementary feed in winter. During the period from September to April animals are weighed 5–6 times. In September, body masses of the focal females did not differ from the rest of the herd. Heart rate (HR) and subcutaneous body temperature (Tsc) data In February 2018, the focal females were instrumented with a heart rate (HR) and temperature logger (DST centi-HRT, Star-Oddi, Gardabaer, Iceland). The surgical protocol is described in the Supporting Information. The DST centi-HRT sensors recorded HR and subcutaneous body temperature (Tsc) every 15 min. HR was automatically calculated from a 4-sec electrocardiogram (ECG) at 150 Hz measurement frequency, alongside an index for signal quality. Additional data processing is described in Supporting Information. Activity data The animals were fitted with collar-mounted tri-axial accelerometers (Vertex Plus Activity Sensor, Vectronic Aerospace GmbH, Berlin, Germany) to monitor their activity levels. These sensors recorded acceleration (g) in three directions representing back-forward, lateral, and dorsal-ventral movements at 8 Hz resolution. For each axis, partial dynamic body acceleration (PDBA) was calculated by subtracting the static acceleration using a 4 sec running average from the raw acceleration (Shepard et al., 2008). We estimated vectorial dynamic body acceleration (VeDBA) by calculating the square root of the sum of squared PDBAs (Wilson et al., 2020). We aggregated VeDBA data into 15-min sums (hereafter “sum VeDBA”) to match with HR and Tsc records. Corrections for time offsets are described in Supporting Information. Due to logger failures, only 10 of the 14 individuals had complete data from both loggers (activity and heart rate). Weather and climate data We set up a HOBO weather station (Onset Computer Corporation, Bourne, MA, USA) mounted on a 2 m tall tripod in May 2018 that measured air temperature (Ta, °C) at 15-minute intervals. The placement of the station was between the two summer paddocks. These measurements were matched to the nearest timestamps for VeDBA, HR and Tsc recordings. Also, we obtained weather records from the nearest public weather stations for the years 1990–2021 (Table S2). Weather station IDs and locations relative to the study area are shown in Figure S1 in the Supporting Information. The temperatures at the study site and the nearest weather station were strongly correlated (Pearson’s, r = 0.99), but temperatures were on average ~1.0°C higher at the study site (Figure S2). Statistical analyses All statistical analyses were conducted in R version 4.1.1 (The R Core Team, 2021). Mean values are presented with standard deviation (SD), and parameter estimates with standard error (SE). Environmental effects on activity states and transition probabilities We fitted hidden Markov models (HMM) to 15-min sum VeDBA using the package ‘momentuHMM’ (McClintock & Michelot, 2018). HMMs assume that the observed pattern is driven by an underlying latent state sequence (a finite Markov chain). These states can then be used as proxies to interpret the animal’s unobserved behaviour (Langrock et al., 2012). We assumed only two underlying states, thought to represent ‘inactive’ and ‘active’ (Figure S3). The ‘active’ state thus contains multiple forms of movement, e.g., foraging, walking, and running, but reindeer spend more than 50% of the time foraging in summer (Skogland, 1980). We fitted several HMMs to evaluate both external (temperature and time of day) and individual-level (calf status) effects on the probability to occupy each state (stationary state probabilities). The combination of the explanatory variables in each HMM is listed in Table S5. Ta was fitted as a continuous variable with piecewise polynomial spline with 8 knots, asserted from visual inspection of the model outputs. We included sine and cosine terms for time of day to account for cyclicity. In addition, to assess the impact of Ta on activity patterns, we fitted five temperature-day categories in interaction with time of day. These categories were based on 20% intervals of the distribution of temperature data from our local weather station, in the period 19 June to 19 August 2018, with ranges of < 10°C (cold), 10−13°C (cool), 13−16°C (intermediate) 16−20°C (warm) and ≥ 20°C (hot). We evaluated the significance of each variable on the transition probabilities from the confidence intervals of each estimate, and the goodness-of-fit of each model using Akaike information criteria (AIC) (Burnham & Anderson, 2002), retaining models within ΔAIC < 5. We extracted the most likely state occupied by an individual using the viterbi function, returning the optimal state pathway, i.e., a two-level categorical variable indicating whether the individual was most likely resting or active. We used this output to calculate daily activity budgets (% time spent active). Drivers of heart rate (HR) and subcutaneous body temperature (Tsc) We matched the activity states derived from the HMM to the HR and Tsc data. We opted to investigate the drivers of variation in HR and Tsc only within the inactive state. HR and Tsc were fitted as response variables in separate generalised additive mixed-effects models (GAMM), which included the following smooth terms: calendar day as a thin-plate regression spline, time of day (ToD, in hours, knots [k] = 10) as a cubic circular regression spline and individual as random intercept. All models were fitted using restricted maximum likelihood, a penalization value (λ) of 1.4 (Wood, 2017), and an autoregressive structure (AR1) to account for temporal autocorrelation. We used the ‘gam.check’ function from the ‘mgcv’ package to select k. The sum of VeDBA in the past 15 minutes was included as a predictor in all models. All models were fitted with the same set of explanatory variables: sum VeDBA, age, body mass (BM), lactation status, Ta, as well as the interaction between lactation status and Ta. Description of files 1. Data: "kutuharju_weather.csv" weather data recorded from local weather station during study period "Inari_Ivalo_lentoasema.csv" public weather data from weather station ID 102033, owned and managed by the Finnish Meterorological Institute "activitydata.Rdata" dataset used in analyses of activity patterns in reindeer "HR_temp_data.Rdata" dataset used in analyses of heart rate and body temperature responses in reindeer "HRfigureData.Rdata" and "TempFigureData.Rdata" are data files (lists) with model outputs generated in "heartrate_bodytemp_analyses.R" and used in "figures_in_paper.R" "HMM_df_withStates.Rdata" data frame used in HMM models including output from viterbi function "plotdf_m16.Rdata" dataframe for plotting output from model 16 "plotdf_m22.Rdata" dataframe for plotting output from model 22 2. Scripts "activitydata_HMMs.R" R script for data prep and hidden markov models to analyse activity patterns in reindeer "heartrate_bodytemp_analyses.R" R script for data prep and generalized additive mixed models to analyse heart rate and body temperature responses in reindeer "figures_in_paper.R" R script for generating figures 1-3 in the manuscript 3. HMM_model "modelList.Rdata" list containing 2 items: string of all 25 HMM models created, and dataframe with model number and formula "m16.Rdata" and "m22.Rdata" direct acces to two best-fit models

The Russian Financial Statements Database (RFSD) is an open, harmonized collection of annual unconsolidated financial statements of the universe of Russian firms:

• 🔓 First open data set with information on every active firm in Russia.

• 🗂️ First open financial statements data set that includes non-filing firms.

• 🏛️ Sourced from two official data providers: the Rosstat and the Federal Tax Service.

• 📅 Covers 2011-2023 initially, will be continuously updated.

• 🏗️ Restores as much data as possible through non-invasive data imputation, statement articulation, and harmonization.

The RFSD is hosted on 🤗 Hugging Face and Zenodo and is stored in a structured, column-oriented, compressed binary format Apache Parquet with yearly partitioning scheme, enabling end-users to query only variables of interest at scale.

The accompanying paper provides internal and external validation of the data: http://arxiv.org/abs/2501.05841.

Here we present the instructions for importing the data in R or Python environment. Please consult with the project repository for more information: http://github.com/irlcode/RFSD.

Importing The Data

You have two options to ingest the data: download the .parquet files manually from Hugging Face or Zenodo or rely on 🤗 Hugging Face Datasets library.

Python

🤗 Hugging Face Datasets

It is as easy as:

from datasets import load_dataset import polars as pl

This line will download 6.6GB+ of all RFSD data and store it in a 🤗 cache folder

RFSD = load_dataset('irlspbru/RFSD')

Alternatively, this will download ~540MB with all financial statements for 2023# to a Polars DataFrame (requires about 8GB of RAM)

RFSD_2023 = pl.read_parquet('hf://datasets/irlspbru/RFSD/RFSD/year=2023/*.parquet')

Please note that the data is not shuffled within year, meaning that streaming first n rows will not yield a random sample.

Local File Import

Importing in Python requires pyarrow package installed.

import pyarrow.dataset as ds import polars as pl

Read RFSD metadata from local file

RFSD = ds.dataset("local/path/to/RFSD")

Use RFSD_dataset.schema to glimpse the data structure and columns' classes

print(RFSD.schema)

Load full dataset into memory

RFSD_full = pl.from_arrow(RFSD.to_table())

Load only 2019 data into memory

RFSD_2019 = pl.from_arrow(RFSD.to_table(filter=ds.field('year') == 2019))

Load only revenue for firms in 2019, identified by taxpayer id

RFSD_2019_revenue = pl.from_arrow( RFSD.to_table( filter=ds.field('year') == 2019, columns=['inn', 'line_2110'] ) )

Give suggested descriptive names to variables

renaming_df = pl.read_csv('local/path/to/descriptive_names_dict.csv') RFSD_full = RFSD_full.rename({item[0]: item[1] for item in zip(renaming_df['original'], renaming_df['descriptive'])})

R

Local File Import

Importing in R requires arrow package installed.

library(arrow) library(data.table)

Read RFSD metadata from local file

RFSD <- open_dataset("local/path/to/RFSD")

Use schema() to glimpse into the data structure and column classes

schema(RFSD)

Load full dataset into memory

scanner <- Scanner$create(RFSD) RFSD_full <- as.data.table(scanner$ToTable())

Load only 2019 data into memory

scan_builder <- RFSD$NewScan() scan_builder$Filter(Expression$field_ref("year") == 2019) scanner <- scan_builder$Finish() RFSD_2019 <- as.data.table(scanner$ToTable())

Load only revenue for firms in 2019, identified by taxpayer id

scan_builder <- RFSD$NewScan() scan_builder$Filter(Expression$field_ref("year") == 2019) scan_builder$Project(cols = c("inn", "line_2110")) scanner <- scan_builder$Finish() RFSD_2019_revenue <- as.data.table(scanner$ToTable())

Give suggested descriptive names to variables

renaming_dt <- fread("local/path/to/descriptive_names_dict.csv") setnames(RFSD_full, old = renaming_dt$original, new = renaming_dt$descriptive)

Use Cases

🌍 For macroeconomists: Replication of a Bank of Russia study of the cost channel of monetary policy in Russia by Mogiliat et al. (2024) — interest_payments.md

🏭 For IO: Replication of the total factor productivity estimation by Kaukin and Zhemkova (2023) — tfp.md

🗺️ For economic geographers: A novel model-less house-level GDP spatialization that capitalizes on geocoding of firm addresses — spatialization.md

FAQ

Why should I use this data instead of Interfax's SPARK, Moody's Ruslana, or Kontur's Focus?hat is the data period?

To the best of our knowledge, the RFSD is the only open data set with up-to-date financial statements of Russian companies published under a permissive licence. Apart from being free-to-use, the RFSD benefits from data harmonization and error detection procedures unavailable in commercial sources. Finally, the data can be easily ingested in any statistical package with minimal effort.

What is the data period?

We provide financials for Russian firms in 2011-2023. We will add the data for 2024 by July, 2025 (see Version and Update Policy below).

Why are there no data for firm X in year Y?

Although the RFSD strives to be an all-encompassing database of financial statements, end users will encounter data gaps:

We do not include financials for firms that we considered ineligible to submit financial statements to the Rosstat/Federal Tax Service by law: financial, religious, or state organizations (state-owned commercial firms are still in the data).

Eligible firms may enjoy the right not to disclose under certain conditions. For instance, Gazprom did not file in 2022 and we had to impute its 2022 data from 2023 filings. Sibur filed only in 2023, Novatek — in 2020 and 2021. Commercial data providers such as Interfax's SPARK enjoy dedicated access to the Federal Tax Service data and therefore are able source this information elsewhere.

Firm may have submitted its annual statement but, according to the Uniform State Register of Legal Entities (EGRUL), it was not active in this year. We remove those filings.

Why is the geolocation of firm X incorrect?

We use Nominatim to geocode structured addresses of incorporation of legal entities from the EGRUL. There may be errors in the original addresses that prevent us from geocoding firms to a particular house. Gazprom, for instance, is geocoded up to a house level in 2014 and 2021-2023, but only at street level for 2015-2020 due to improper handling of the house number by Nominatim. In that case we have fallen back to street-level geocoding. Additionally, streets in different districts of one city may share identical names. We have ignored those problems in our geocoding and invite your submissions. Finally, address of incorporation may not correspond with plant locations. For instance, Rosneft has 62 field offices in addition to the central office in Moscow. We ignore the location of such offices in our geocoding, but subsidiaries set up as separate legal entities are still geocoded.

Why is the data for firm X different from https://bo.nalog.ru/?

Many firms submit correcting statements after the initial filing. While we have downloaded the data way past the April, 2024 deadline for 2023 filings, firms may have kept submitting the correcting statements. We will capture them in the future releases.

Why is the data for firm X unrealistic?

We provide the source data as is, with minimal changes. Consider a relatively unknown LLC Banknota. It reported 3.7 trillion rubles in revenue in 2023, or 2% of Russia's GDP. This is obviously an outlier firm with unrealistic financials. We manually reviewed the data and flagged such firms for user consideration (variable outlier), keeping the source data intact.

Why is the data for groups of companies different from their IFRS statements?

We should stress that we provide unconsolidated financial statements filed according to the Russian accounting standards, meaning that it would be wrong to infer financials for corporate groups with this data. Gazprom, for instance, had over 800 affiliated entities and to study this corporate group in its entirety it is not enough to consider financials of the parent company.

Why is the data not in CSV?

The data is provided in Apache Parquet format. This is a structured, column-oriented, compressed binary format allowing for conditional subsetting of columns and rows. In other words, you can easily query financials of companies of interest, keeping only variables of interest in memory, greatly reducing data footprint.

Version and Update Policy

Version (SemVer): 1.0.0.

We intend to update the RFSD annualy as the data becomes available, in other words when most of the firms have their statements filed with the Federal Tax Service. The official deadline for filing of previous year statements is April, 1. However, every year a portion of firms either fails to meet the deadline or submits corrections afterwards. Filing continues up to the very end of the year but after the end of April this stream quickly thins out. Nevertheless, there is obviously a trade-off between minimization of data completeness and version availability. We find it a reasonable compromise to query new data in early June, since on average by the end of May 96.7% statements are already filed, including 86.4% of all the correcting filings. We plan to make a new version of RFSD available by July.

Licence

Creative Commons License Attribution 4.0 International (CC BY 4.0).

Copyright © the respective contributors.

Citation

Please cite as:

@unpublished{bondarkov2025rfsd, title={{R}ussian {F}inancial {S}tatements {D}atabase}, author={Bondarkov, Sergey and Ledenev, Victor and Skougarevskiy, Dmitriy}, note={arXiv preprint arXiv:2501.05841}, doi={https://doi.org/10.48550/arXiv.2501.05841}, year={2025}}

Acknowledgments and Contacts

Data collection and processing: Sergey Bondarkov, sbondarkov@eu.spb.ru, Viktor Ledenev, vledenev@eu.spb.ru

Project conception, data validation, and use cases: Dmitriy Skougarevskiy, Ph.D.,