This dataset is maintained by Steven Firth (s.k.firth@lboro.ac.uk), Building Energy Research Group (BERG), School of Civil and Building Engineering, Loughborough University. The REFIT project (www.refitsmarthomes.org) carried out a study from 2013 to 2015 in which 20 UK homes were upgraded to Smart Homes through the installation of devices including Smart Meters, programmable thermostats, programmable radiator valves, motion sensors, door sensors and window sensors.Data was collected using building surveys, sensor placements and household interviews.The REFIT Smart Home dataset is one of the datasets made publically available by the project. This dataset includes: - Building survey data for the 20 homes. - Sensor measurements made before the Smart Home equipment was installed. - Sensor measurements made after the Smart Home equipment was installed. - Climate data recorded at a nearby weather station.--- This work has been carried out as part of the REFIT project (‘Personalised Retrofit Decision Support Tools for UK Homes using Smart Home Technology’, Grant Reference EP/K002457/1). REFIT is a consortium of three universities - Loughborough, Strathclyde and East Anglia - and ten industry stakeholders funded by the Engineering and Physical Sciences Research Council (EPSRC) under the Transforming Energy Demand in Buildings through Digital Innovation (BuildTEDDI) funding programme. For more information see: www.epsrc.ac.uk and www.refitsmarthomes.org---The references below provide links to the REFIT project website, the TEDDINET website, a journal article which uses the dataset, and three additional datasets collected as part of the REFIT project by the University of Strathclyde and the University of East Anglia.

This dataset contains qualitative data collected using semi-structured interviews and a structured survey at four time points during the REFIT field trial of smart home technologies which involved 20 households. A related data collection containing survey data is available via Related Resources /Related Data collections. This was also part of the part of the REFIT project. A national survey was conducted to measure perceptions of smart homes. The survey instrument was developed and tested by the project team. The survey was implemented online during September - October 2015 by a market research company using a representative sample of UK homeowners. A total of 1054 responses were collected.

These datasets were collected as part of the REFIT project (`Personalised Retrofit Decision Support Tools for UK Homes using Smart Home Technology’). The REFIT project ran from 2012 - 2015 as a consortium of three universities - Loughborough, Strathclyde and East Anglia - and ten industry stakeholders.

During the REFIT project, twenty households were recruited into a field trial of smart home technologies in Loughborough, UK. The field trial ran from October 2013 - October 2015. Smart home technologies were installed in participating households from May - August 2014.

During the field trial, the REFIT project team collected quantitative data on real-time electricity and gas usage (from smart meters) and qualitative data on perceptions and usage of smart home technologies. This dataset contains the qualitative data.

The qualitative data were collected using semi-structured interviews, surveys, and video ethnography. The video files are not shareable (as they cannot be anonymised). This dataset contains transcripts, notes and responses from the semi-structured interviews and surveys.

The interview and survey data were collected at four time points or cross-sections during the two year field trial. All twenty households participated at time point one (on agreeing to participate in the field trial) and time point two (prior to installation of the smart home technologies). A subset of ten households participated at time point three (immediately after installation of the smart home technologies) and time point four (after installation of the smart home technologies).

Dataset from a nationally representative sample of 2,101 participants in the United Kingdom.

Participants were recruited through the online market research sample aggregator Qualtrics . The quotas for the sample were based on 2016 European statistical datasets for population by age and gender, and educational attainment level.

Age values:

Dataset

This dataset was created by Kaggle@ck01

Released under MIT

Contents

Load Forecasting Dataset

Readme File

VARLAB – The Centre for Research & Technology, Hellas [CERTH] - Informatics and Telematics Institute [ITI] - https://varlab.iti.gr/

Authors: Chrysovalantis-George Kontoulis, Georgios Stavropoulos, Dimosthenis Ioannidis

Publication Date: February -, 2023

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreements No. 957406 (TERMINET).

1.Introduction

This dataset features information from a smarthome located at Greece, which features the Mediterranean climate. The building is utilized as a modern workplace that is being used for various every day activities. It is equipped with numerous smart devices and appliances, from smart lights to smart a elevator, while also featuring PVTs.

2.Dataset Overview

2.1Dataset Collection

The system is built on multiple communication protocols including EnOcean, Zigbee, Modbus, BACnet, and, LTE/IEEE 802.15.4 at 2.4GHz. For the sensor data collection, a raspberry Pi microcontroller was used, and data were subsequently transmitted to the storage database.

The extraction period of the data is between 2021-01-01 through 2022-12-20. Along this period there is a total of 66619 unique recordings and the time granularity of the data is set to 15 minutes for all devices.

2.2Data Peculiarities

The place is occupied from Monday to Friday from 9:00 AM GMT+2 (Greenwich Mean Time) all the way through 5:00 PM GMT+2. Note that the building is not active during Greek public holidays, but some computers or servers might be on and consuming electrical energy. Also, there are some irregularities in the data reporting consistency at summer, Christmas & Easter as the building is not occupied for a long time of period. Timestamps of the dataset are in the GMT+2 timezone.

2.3Dataset Structure

This dataset includes a total of six features and it can be used for Electrical, Thermal and Cooling Load forecasting. Electricity Consumption is the consumption of the whole house, Air-condition Status is either 1 or 0 for on and off, respectively, Luminance is how bright a space is, Light Dimming is the dimming of the lights in each room. Finally we have the Indoor Temperature for each room and the Outdoor Temperature.

Data are extracted from four rooms in total. Note that in rooms 1 and 3, there is only one indoor temperature device, thus values are identical for temperature_room_1 and temperature_room_3. Note that sensors have some null values, which is generally either due to inactivity, e.g., the Light Dimming sensor and the Air-condition Status are event-based or due to potential system downtime.

The provided dataset is stored in csv format. A brief overview of the dataset is presented at the Table 3.1.

Table 2.1 Dataset overview

Censor

Symbolic Naming

Measurement Unit

Electricity Consumption

KWh_S_total

kWh

Air-condition Status

status_room_0

status_room_1

status_room_2

status_room_3

-

Luminance

luminance_room_0

luminance_room_1

luminance_room_2

luminance_room_3

Lux

Light Dimming

dimming_room_0

dimming_room_1

dimming_room_2

dimming_room_3

%

Indoor Temperature

temperature_room_0

temperature_room_1

temperature_room_2

temperature_room_3

°C

Outdoor Temperature

airTemperature

°C

2.4Descriptive Statistics

Table 2.2 provides a brief overview of the key statistical characteristics of the data to. The table presents a summary of important metrics and measures, including measure of central tendency such as the mean, as well as measures of variability such as the standard deviation.

Table 2.2 Descriptive Statistics

Symbolic Naming

Values Count

Mean

Std

Min

Max

KWh_S_total

62877

71511,16

52235,30

2,22

135494,70

status_room_0

status_room_1

status_room_2

status_room_3

16689

14357

13302

13388

0,38

0,15

0,27

0,26

0,49

0,36

0,44

0,44

0,00

0,00

0,00

0,00

1,00

1,00

1,00

1,00

luminance_room_0

luminance_room_1

luminance_room_2

luminance_room_3

31676

14727

6799

23993

169,93

165,95

99.71

205,66

294,60

267,69

157,40

304,14

0.00

0.00

0.00

0.00

1024,00

1024,00

1024,00

1024,00

dimming_room_0

dimming_room_1

dimming_room_2

dimming_room_3

432

683

8

608

1,95

41.29

15,00

42,40

11,23

40.32

22,68

43,43

0,00

0,00

0,00

0, 00

100,00

100,00

50,00

100,00

temperature_room_0

temperature_room_1

temperature_room_2

temperature_room_3

26915

33786

34778

33786

27,50

24,28

23,89

24,28

4,44

2,96

4,52

2,96

17,54

13,95

7,95

13,95

44,30

34,62

35,59

34,62

airTemperature

47647

16,91

8,62

-4,52

40,28

3.Acknowledgment

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreements No. 957406 (TERMINET).

Problem Statement

👉 Download the case studies here

Individuals with disabilities often face challenges in interacting with traditional technology interfaces, limiting their independence and access to essential tools. A technology firm sought to develop a voice-activated system that would enable users with disabilities to control devices, access information, and perform tasks seamlessly using speech commands.

Challenge

Developing a voice recognition system for accessibility required overcoming the following challenges:

Ensuring accurate speech recognition across diverse accents, speech impairments, and environmental noise conditions.

Creating intuitive voice commands that are easy for users to remember and execute.

Integrating the system with a wide range of devices and applications for broad usability.

Solution Provided

A voice-activated accessibility system was developed using speech recognition technology and natural language understanding (NLU). The solution was designed to:

Recognize and interpret user speech commands accurately in real-time.

Perform tasks such as device control, information retrieval, and communication seamlessly.

Provide a user-friendly interface that adapts to individual speech patterns and preferences.

Development Steps

Data Collection

Collected diverse voice samples, including recordings from individuals with speech impairments, to train and enhance the speech recognition models.

Preprocessing

Processed audio data to eliminate background noise, normalize speech patterns, and ensure clarity for model training.

Model Development

Built speech recognition models using advanced machine learning techniques to ensure high accuracy in command recognition. Developed natural language understanding (NLU) modules to interpret complex commands and context.

Testing and Validation

Tested the system with real-world scenarios and diverse user groups, ensuring reliability and usability for individuals with varying speech abilities.

Deployment

Integrated the solution with smart home devices, mobile applications, and communication tools, enabling users to control and interact with technology seamlessly.

Continuous Improvement

Implemented a feedback loop for users to report issues and suggest improvements, refining the system’s capabilities over time.

Results

Enhanced Accessibility

The voice-activated system provided users with disabilities greater access to technology, allowing them to perform tasks independently.

Improved User Interaction

The intuitive voice interface enabled smooth and natural interactions, reducing frustration and enhancing the user experience.

Increased Independence

Users gained the ability to control devices, access information, and communicate without assistance, promoting self-reliance.

Adaptable Technology

The system learned and adapted to individual speech patterns, ensuring consistent performance across diverse user groups.

Broad Integration

Seamless compatibility with smart home devices, smartphones, and other technologies expanded the system’s usability and impact.

This anonymized dataset holds answers of 305 Dutch parents who have at least one child between 3-8 years as well as a Google Assistant-powered smart speaker at home.

This dataset contains transcripts from semi-structured interviews with 15 Danish households with smart home technology installed. The interviews were conducted in 2020 as part of the eCAPE project (New energy consumer roles and smart technologies – actors, practices and equality) running from 2018-2023. eCAPE is financed by the European Research Council (ERC) under the European Union´s Horizon 2020 research and innovation program (grant agreement number 786643). The project is led by Professor Kirsten Gram-Hanssen from Department of the Built Environment, Aalborg University, and the interviews were conducted by Line Kryger Aagaard from Department of the Built Environment, Aalborg University, as part of her PhD study within eCAPE. Interviews were transcribed by Line Kryger Aagaard and a student assistant.

The participating households had a combination of various smart home technologies (e.g. smart lighting, smart heating, digital voice assistants, robotic vacuum cleaners). The recruitment process is described in a paper by Aagaard & Madsen (2022): “The first nine households in the study were recruited via SHT [smart home technology] Facebook groups where people share experiences and advice, and six additional households were recruited via snowball sampling referred from the initial participants and from one contact of the authors. Only men responded to the posted research call and common to all of them was an outspoken interest in technology. They all lived in opposite-sex relationships, except one single man, and were asked to bring their female partners for the interviews, which 12 out of 14 did” (Aagaard & Madsen 2022: 680).

All names are pseudonyms and participants’ age, occupation and other personal data have in some cases been altered and in some cases paragraphs have been removed, all to ensure anonymity. Alterations are marked with bold writing and removed parts are indicated with brackets and X’s. Participants received written and oral information on the research purpose and handling of their personal data and gave their written consent to participate. The interviews were conducted in the participants’ homes, except from one interview that was conducted online. Photos were taken during the interviews with participants’ consent, but these photos have not been published due to the protection of the anonymity of participants.

Apart from the interview transcripts this dataset contains the interview guide (Danish and English version) and the posted call for research participants (Danish and English version).

To this date, the interview data have been analyzed in two journals papers:

Aagaard, Line Kryger. 2022. ‘When Smart Technologies Enter Household Practices: The Gendered Implications of Digital Housekeeping’. Housing, Theory and Society 0 (0): 1–18. https://doi.org/10.1080/14036096.2022.2094460.

Aagaard, Line Kryger, and Line Valdorff Madsen. 2022. ‘Technological Fascination and Reluctance: Gendered Practices in the Smart Home’. Buildings and Cities 3 (1): 677–91. https://doi.org/10.5334/bc.205.

This dataset is part of the UCR Archive maintained by University of Southampton researchers. Please cite a relevant or the latest full archive release if you use the datasets. See http://www.timeseriesclassification.com/.

The collection of this data was part of the project titled Personalised Retrofit Decision Support Tools for UK Homes using Smart Home Technology (REFIT). The REFIT dataset includes data from 20 households from Loughborough area over the period 2013-2014 (see [1]). We use data of freezers in House 1 to make two datasets. These two datasets share a same test set and only differ in the number of training instances. There are two classes, one representing the power demand of the fridge freezer in the kitchen, the other representing the power demand of the (less frequently used) freezer in the garage. They are hard to tell apart globally but they differ locally.

[1] Murray, David et al., "A data management platform for personalised real-time energy feedback", Proceedings of the 8th International Conference on Engery Efficiency in Domestic Appliances and Lighting, 2015.

Donator: REFIT project

"Smart meter roll-outs provide easy access to granular meter measurements, enabling advanced energy services, ranging from demand response measures, tailored energy feedback and smart home/building automation. To design such services, train and validate models, access to data that resembles what is expected of smart meters, collected in a real-world setting, is necessary. The REFIT electrical load measurements dataset includes whole house aggregate loads and nine individual appliance measurements at 8-second intervals per house, collected continuously over a period of two years from 20 houses. During monitoring, the occupants were conducting their usual routines. At the time of publishing, the dataset has the largest number of houses monitored in the United Kingdom at less than 1-minute intervals over a period greater than one year. The dataset comprises 1,194,958,790 readings, that represent over 250,000 monitored appliance uses. The data is accessible in an easy-to-use comma-separated format, is time-stamped and cleaned to remove invalid measurements, correctly label appliance data and fill in small gaps of missing data."

Data set: https://pureportal.strath.ac.uk/en/publications/an-electrical-load-measurements-dataset-of-united-kingdom-househo

Banner photo by @thejmoore on unsplash.com

jinaai_jina-embeddings-v2-base-en-2024615-ioyu-webapp Dataset

  Dataset Description

The dataset "Smart home technology brand" is a generated dataset designed to support the development of domain specific embedding models for retrieval tasks.

  Associated Model

This dataset was used to train the jinaai_jina-embeddings-v2-base-en-2024615-ioyu-webapp model.

  How to Use

To use this dataset for model training or evaluation, you can load it using the Hugging… See the full description on the dataset page: https://huggingface.co/datasets/fine-tuned/jinaai_jina-embeddings-v2-base-en-2024615-ioyu-webapp.

We hypothesize cognitive impairment can be evident in everyday task performance. We also postulate that differences in task performance can be automatically detected between cognitively healthy individuals and those with dementia and mild cognitive impairment (MCI) using smart home and ubiquitous computing technologies. This dataset contains ambient sensor readings collected in the CASAS smart apartment testbed at Washington State University for 179 participants with corresponding cognitive diagnoses.

Data are collected continuously from ambient sensors while participants perform 24 scripted activities. Each sensor reading is reported on a separate line and is described by fields date, time, sensor, and message. The first 8 activities are performed without cues (task step reminders), the second 8 activities are performed with cues when needed, and the last 8 activities, part of a complex day out task, are interwoven in a natural manner without cues. The task list is included in the file activitylist.txt.

The file activityscores.txt provides numeric values for activities 1-8 that are based on experimenter assessment of task quality. The file also gives scores for the day out task. Finally, the file diagnosis.txt lists the diagnosis for each participant, coded as the following:

1 = dementia
2 = MCI
3 = middle age 45-59
4 = young-old 60-74
5 = old-old 75+
6 = other medical
7 = watch/at risk - follow longitudinally
8 = younger adult
9 = younger adult, English second language
10 = diagnosis not available

The sensors are categorized (and named) as:

• M01 - M51: PIR motion detectors (ON when detected motion starts and OFF when it stops)
• I01 - I10: item use sensors (PRESENT or ABSENT indicating item is on sensor or not)
• D01 - D019: door sensor on cabinets and doors (OPEN or CLOSE)
• P001 and P002: current eletricity consumption
• T001 - T006: ambient temperature sensors
• BATP and BATV: sensor battery levels

Description

This qualitative dataset is the first part of a PhD study on co-designing smart home technologies, and represents the data collected during a series of independent in-person workshops with professionals developing smart technology, its early-adopters, and late/non-adopters. The data collected during the subsequent parts of the referred study are also available at Zenodo.

Documents from workshop with professionals

P1_WSP-PRO-TRANSCR_R02.docx (transcription of the workshop's audio recordings)

P1_WSP-PRO-VIS_000 till _013.jpg (participant-generated visual data)

Documents from workshop with early-adopters

P1_WSP-EA-TRANSCR_R01.docx (transcription of the workshop's audio recordings)

P1_WSP-EA-VIS_000 till _011.jpg (participant-generated visual data)

Documents from workshop with late/non-adopters

P1_WSP-LN-TRANSCR_R00.docx (transcription of the workshop's audio recordings)

P1_WSP-LN-VIS_000 till _012.jpg (participant-generated visual data)

Acknowledgements

This study is part of the GECKO Project (https://gecko-project.eu/) and has received funding from the European Commission under the Horizon2020 MSCA-ITN-2020 Innovative Training Networks programme, Grant Agreement No 955422 (https://cordis.europa.eu/project/id/955422).

Percentage of Internet users by selected Internet service and technology, such as; home Internet access, use of smart home devices, use of smartphones, use of social networking accounts, use or purchase of streaming services, use of government services online and online shopping.

Abstract: The dataset contains electrical consumption data from a three months living lab experiment in the Energy Smart Home Lab (ESHL) at Karlsruhe Institute of Technology. During the experiment two tenants lived their daily life in the ESHL, a smart home consisting of a 60m² living area and a 20m² technical room. The dataset contains the electrical consumption data for every device and socket in the living area as well as the technical room. There is one set for the average electricity demand per 15min for all devices in the living room, one set for the average demand per 15min for all devices in the technical room, one set for the maximum demand per 15min for the devices in the living room and one set for the maximum demand per 15min for the devices in the technical room. Due to technical difficulties the recording for the devices in the technical room started two weeks after the recording of the living area. There are also some shorter gaps in the data due to power outages. In addition to the electrical consumption data, a timeseries for the outdoor temperature with a temporal resolution of 15min is included in the dataset.

This dataset contains transcripts from 8 semi-structured qualitative interviews, conducted as part of the research project eCAPE (New Energy Consumer Roles and Technologies – Actors, Practices and Equality), running from 2018-2024. eCAPE is financed by the European Research Council (ERC) under the European Union´s Horizon 2020 research and innovation program (grant agreement number 786643). The project is led by Professor Kirsten Gram-Hanssen from Department of the Built Environment, Aalborg University.

This sub-project examined how consumers and residents are part of a flexible energy system using smart heating technologies, and how such technologies affect their everyday lives. This part of the project mainly deals with everyday life at home, daily routines and tasks, energy consumption and comfort, as well as interactions with technologies and appropriation.

The interviews were conducted in 2019-20 and different types of households were interviewed in their homes, on different locations. The interviews were conducted by Line Valdorff Madsen from Department of the Built Environment, Aalborg University. Interviews were transcribed by a professional. The interviews were coded and analysed in Nvivo by Line Valdorff Madsen. The recruitment process and interview method are further described in the related peer-reviewed publications.

All names are removed in transcripts and participants’ occupation and other personal data have in some cases been altered or removed to ensure anonymity. Alterations are marked with brackets, where anoymised words or sections are indicated with [anonymised].

Participants received written and oral information on the research purpose and handling of their personal data and gave their written or oral consent to participate. The interviews were conducted in the participants’ homes, except for one telephone interview. Photos were taken during the interviews with participants’ consent, but these photos are not published in the dataset due to the protection of the anonymity of participants. Apart from the interview transcripts this dataset contains the interview guide (Danish and English version).

Apart from the interview transcripts this dataset contains the interview guide (Danish and English version) and the posted call for research participants (Danish and English version).

To this date, the interview data have been analyzed in two journal papers:

• Aagaard, L. K., & Madsen, L. V. 2022. ‘Technological Fascination and Reluctance: Gendered Practices in the Smart Home’. Buildings and Cities 3 (1): 677–91. https://doi.org/10.5334/bc.205.
• Madsen, L. V., Larsen, S. P. A. K. & Hansen, A. R. 2023. 'Embodied competencies and smart home technology in energy use: three ways users integrate smart heating systems in everyday practices'. Energy Efficiency 16 (55): https://doi.org/10.1007/s12053-023-10138-0

The dataset refers to measurements and data collected from and for the CERTH Smart House nanogrid infrastructure in Thessaloniki, Greece. This infrastructure is a living lab that belongs to the Centre for Research and Technology-Hellas (www.certh.gr) and has been designed, deployed and operated by the Information Technology Institute (www.iti.gr). Mainly energy-related aspects are included in the datasets uploaded which are in the form of .csv files, covering Electricity Energy Consumption, Generation, and Storage, as well as Weather and Electricity Price information from external APIs (a local weather station has just been installed and will be included in future versions).

This dataset underlies the analysis reported on in a comment published by ESC: https://es.catapult.org.uk/insight/spotting-low-carbon-technologies-charlotte-avery/. The purpose of this analysis was to create a simple and easy method of identifying low carbon technologies from smart meter alone. Dataset: - Half-hourly electric smart meter readings during winter months (Jan/Feb 2023) for just under 300 real-world homes in the UK via integrations with ESC's Living Lab. - Qualitative data indicating the presence of a low carbon technology (EV or heat pump) in each home for these homes determined by questionnaire/survey responses via integrations with ESC's Living Lab.

The global Artificial Intelligence (AI) Training Dataset market is experiencing robust growth, driven by the increasing adoption of AI across diverse sectors. The market's expansion is fueled by the burgeoning need for high-quality data to train sophisticated AI algorithms capable of powering applications like smart campuses, autonomous vehicles, and personalized healthcare solutions. The demand for diverse dataset types, including image classification, voice recognition, natural language processing, and object detection datasets, is a key factor contributing to market growth. While the exact market size in 2025 is unavailable, considering a conservative estimate of a $10 billion market in 2025 based on the growth trend and reported market sizes of related industries, and a projected CAGR (Compound Annual Growth Rate) of 25%, the market is poised for significant expansion in the coming years. Key players in this space are leveraging technological advancements and strategic partnerships to enhance data quality and expand their service offerings. Furthermore, the increasing availability of cloud-based data annotation and processing tools is further streamlining operations and making AI training datasets more accessible to businesses of all sizes. Growth is expected to be particularly strong in regions with burgeoning technological advancements and substantial digital infrastructure, such as North America and Asia Pacific. However, challenges such as data privacy concerns, the high cost of data annotation, and the scarcity of skilled professionals capable of handling complex datasets remain obstacles to broader market penetration. The ongoing evolution of AI technologies and the expanding applications of AI across multiple sectors will continue to shape the demand for AI training datasets, pushing this market toward higher growth trajectories in the coming years. The diversity of applications—from smart homes and medical diagnoses to advanced robotics and autonomous driving—creates significant opportunities for companies specializing in this market. Maintaining data quality, security, and ethical considerations will be crucial for future market leadership.

Here are a few use cases for this project:

1. Security Surveillance: The "Company" model can be used in a security surveillance system where it identifies and logs individuals detected in the footage, helping to maintain safe environments in both public and private settings.

2. Attendance Management: For office environments or events, the model could be used to manage attendance by recognizing and recording the entrance and exit of individuals.

3. Retail Analytics: The model could provide valuable insights to retailers about foot traffic, tracking who comes in and out of the store, distinguishing between staff and customer.

4. Interactive Experiences: In museums or educational facilities, it could be used to create interactive experiences where the system identifies the number of people watching an exhibit and personalizes the content accordingly.

5. Smart Home Technology: "Company" model can also be used in smart home technologies for recognizing authorized personnel in a given space to automate certain processes like personalized settings, security alerts, etc.