Measuring the quality of Question Answering (QA) systems is a crucial task to validate the results of novel approaches. However, there are already indicators of a reproducibility crisis as many published systems have used outdated datasets or use subsets of QA benchmarks, making it hard to compare results. We identified the following core problems: there is no standard data format, instead, proprietary data representations are used by the different partly inconsistent datasets; additionally, the characteristics of datasets are typically not reflected by the dataset maintainers nor by the system publishers. To overcome these problems, we established an ontology---Question Answering Dataset Ontology (QADO)---for representing the QA datasets in RDF. The following datasets were mapped into the ontology: the QALD series, LC-QuAD series, RuBQ series, ComplexWebQuestions, and Mintaka. Hence, the integrated data in QADO covers widely used datasets and multilinguality. Additionally, we did intensive analyses of the datasets to identify their characteristics to make it easier for researchers to identify specific research questions and to select well-defined subsets. The provided resource will enable the research community to improve the quality of their research and support the reproducibility of experiments.

Here, the mapping results of the QADO process, the SPARQL queries for data analytics, and the archived analytics results file are provided.

Up-to-date statistics can be created automatically by the script provided at the corresponding QADO GitHub RDFizer repository.

OpenAI's o1 had the highest share of questions answered when attempted in SimpleQA benchmark in 2025. Claude-3 had the highest share of simply not attempting questions, though whether this is due to lack of data or other reasons is unknown.

Full text of questions and answers from Stack Overflow that are tagged with the r tag, useful for natural language processing and community analysis.

This is organized as three tables:

• Questions contains the title, body, creation date, score, and owner ID for each R question.
• Answers contains the body, creation date, score, and owner ID for each of the answers to these questions. The ParentId column links back to the Questions table.
• Tags contains the tags on each question besides the R tag.

For space reasons only non-deleted and non-closed content are included in the dataset. The dataset contains questions up to 24 September 2017 (UTC).

License

All Stack Overflow user contributions are licensed under CC-BY-SA 3.0 with attribution required.

Here are a few use cases for this project:

1. Digital Document Management: This model can be used to effectively organize and manage digital documents. By identifying areas such as headers, addresses, and vendors, it could streamline workflows in companies dealing with large amounts of papers, forms or invoices.

2. Automated Data Extraction: The model could be used in extracting pertinent information from documents automatically. For example, pulling out questions and answers from educational materials, extracting vendor or address information from invoices, or grabbing column headers from statistical reports.

3. Augmented Reality (AR) Applications: "Question Answers Label" can be utilized in AR glasses to give real-time information about objects a user sees, especially in the realm of paper documents.

4. Virtual Assistance: This model may be used to build a virtual assistant capable of reading and understanding physical documents. For instance, reading out a user's mail, helping learning from textbooks, or assisting in reviewing legal documents.

5. Accessibility Tools for Visually Impaired: The tool could be utilized to interpret written documents for visually impaired people by identifying and vocalizing text based on their classes (answers, questions, headers, etc).

We develop a panel data model explaining answers to subjective probabilities about binary events and estimate it using data from the Health and Retirement Study on six such probabilities. The model explicitly accounts for several forms of reporting behavior: rounding, focal point 50% answers and item nonresponse. We find observed and unobserved heterogeneity in the tendencies to report rounded values or a focal answer, explaining persistency in 50% answers over time. Focal 50% answers matter for some of the probabilities. Incorporating reporting behavior does not have a large effect on the estimated distribution of the genuine subjective probabilities.

This dataset contains the data used in the master's thesis with the above title. It consists of a BibTeX file with the bibliographic metadata of the publications and websites cited throughout the thesis, and a Markdown file with statistics of the data sources discussed in Chapter 4.

This statistic represents the trusted place to receive answers on questions of faith among Iraqi Millennials as of 2017. During the survey, ** percent of Iraqi Millennials stated that they would go to their local mosque Imam for answers on their questions of faith.

The average environmental and occupational physiologist may find statistics are difficult to interpret and use since their formal training in statistics is limited. Unfortunately, poor statistical practices can generate erroneous or at least misleading results and distorts the evidence in the scientific literature. These problems are exacerbated when statistics are used as thoughtless ritual that is performed after the data are collected. The situation is worsened when statistics are then treated as strict judgements about the data (i.e., significant versus non-significant) without a thought given to how these statistics were calculated or their practical meaning. We propose that researchers should consider statistics at every step of the research process whether that be the designing of experiments, collecting data, analysing the data or disseminating the results. When statistics are considered as an integral part of the research process, from start to finish, several problematic practices can be mitigated. Further, proper practices in disseminating the results of a study can greatly improve the quality of the literature. Within this review, we have included a number of reminders and statistical questions researchers should answer throughout the scientific process. Rather than treat statistics as a strict rule following procedure we hope that readers will use this review to stimulate a discussion around their current practices and attempt to improve them. The code to reproduce all analyses and figures within the manuscript can be found at https://doi.org/10.17605/OSF.IO/BQGDH.

Both Binomial and Thompson sampling are more efficient than Random sampling (lower 〈A〉) without losing the crowd’s average consensus on answers, measured by 〈S〉 and 〈d〉.

Full text of questions and answers from Cross Validated, the statistics and machine learning Q&A site from the Stack Exchange network.

This is organized as three tables:

• Questions contains the title, body, creation date, score, and owner ID for each question.
• Answers contains the body, creation date, score, and owner ID for each of the answers to these questions. The ParentId column links back to the Questions table.
• Tags contains the tags on each question

For space reasons only non-deleted and non-closed content are included in the dataset. The dataset contains questions up to 19 October 2016 (UTC).

License

All Stack Exchange user contributions are licensed under CC-BY-SA 3.0 with attribution required.

Aggregate statistics of the answer rate of questions in each site.

These datasets contain 1.48 million question and answer pairs about products from Amazon.

Metadata includes

• question and answer text

• is the question binary (yes/no), and if so does it have a yes/no answer?

• timestamps

• product ID (to reference the review dataset)

Basic Statistics:

• Questions: 1.48 million

• Answers: 4,019,744

• Labeled yes/no questions: 309,419

• Number of unique products with questions: 191,185

We make only one point in this article. Every quantitative study must be able to answer the question: what is your estimand? The estimand is the target quantity---the purpose of the statistical analysis. Much attention is already placed on how to do estimation; a similar degree of care should be given to defining the thing we are estimating. We advocate that authors state the central quantity of each analysis---the theoretical estimand---in precise terms that exist outside of any statistical model. In our framework, researchers do three things: (1) set a theoretical estimand, clearly connecting this quantity to theory, (2) link to an empirical estimand, which is informative about the theoretical estimand under some identification assumptions, and (3) learn from data. Adding precise estimands to research practice expands the space of theoretical questions, clarifies how evidence can speak to those questions, and unlocks new tools for estimation. By grounding all three steps in a precise statement of the target quantity, our framework connects statistical evidence to theory.

Google Assistant ranked first among selected digital voice assistants in terms of share of questions answered correctly, as of 2020. A research found that nearly ** percent of the simple questions and over ** percent of the complex questions asked were answered correctly by Google Assistant. Complex questions involved comparison, composition, and/or reasoning.

This statistic represents the importance for obtaining answers on questions of faith among Sudanese Millennials as of 2017. During the survey, ** percent of Sudanese Millennials stated that obtaining an answer for their questions of faith was very important to them.

The solutions of mysteries can lead to salvation for those on the reference desk dealing with business students or difficult questions.

Stack Overflow Programming Questions Dataset (2020-2025)

Overview

This comprehensive dataset contains 95,636 programming questions from Stack Overflow, covering 20 popular programming languages collected over a 5-year period (2020-2025). Each question includes detailed metadata, top answers, and quality metrics.

Dataset Statistics

• Total Questions: 95,636
• Programming Languages: 20
• Time Period: 2020-2025
• Features: 34 columns
• Dataset Size: ~130 MB
• Answer Rate: 54.79%
• Code Presence: 92.62%
• Uniqueness: 99.99%

Programming Languages Included

1. Python (6,491 questions)
2. JavaScript (7,355 questions)
3. Java (5,948 questions)
4. C++ (5,272 questions)
5. C# (5,167 questions)
6. Swift (5,044 questions)
7. R (5,014 questions)
8. C (4,869 questions)
9. Rust (4,847 questions)
10. Ruby (4,846 questions)
11. TypeScript (4,143 questions)
12. Scala (4,526 questions)
13. Kotlin (4,543 questions)
14. Go (4,810 questions)
15. PHP (4,780 questions)
16. MATLAB (4,157 questions)
17. Perl (3,854 questions)
18. HTML (2,891 questions)
19. CSS (1,762 questions)
20. SQL (4,687 questions)

Features

Question Information

• question_id: Unique Stack Overflow question ID
• title: Question title
• body: Full question body (HTML formatted)
• tags: Comma-separated tags
• programming_language: Primary programming language

Metrics

• view_count: Number of views
• score: Question score (upvotes - downvotes)
• answer_count: Number of answers
• is_answered: Whether question has accepted answer
• has_accepted_answer: Whether question has accepted answer

Content Analysis

• has_code: Whether question contains code blocks
• code_block_count: Number of code blocks
• body_word_count: Word count in question body
• body_char_count: Character count in question body
• title_word_count: Word count in title

Quality Metrics

• difficulty_score: Calculated difficulty score (0-1)
• quality_score: Calculated quality score (0-1)
• owner_reputation: Question owner's reputation

Temporal Features

• creation_date: Question creation timestamp
• creation_year: Year of creation
• creation_month: Month of creation
• creation_weekday: Day of week (0=Monday)
• last_activity_date: Last activity timestamp
• first_response_time_seconds: Time to first answer (seconds)

Answer Information

• top_answer_score: Score of top answer
• top_answer_body_length: Length of top answer body
• accepted_answer_score: Score of accepted answer

Data Collection Methodology

• Source: Stack Exchange API (official API)
• Collection Period: November 2020 - November 2025
• Filters Applied:
  • Minimum 100 views
  • Minimum 1 answer
  • Questions with body content
• Answer Collection: Top 3 answers per question
• Data Cleaning: Duplicate removal, HTML cleaning, validation

Use Cases

1. Natural Language Processing (NLP)

   • Question classification
   • Sentiment analysis
   • Topic modeling
   • Text generation

2. Machine Learning

   • Question quality prediction
   • Answer recommendation systems
   • Duplicate question detection
   • Difficulty estimation

3. Data Science Research

   • Programming language trends
   • Developer behavior analysis
   • Community engagement patterns
   • Technical knowledge evolution

4. Educational Applications

   • Learning resource generation
   • Difficulty assessment
   • Curriculum development
   • Student assessment tools

5. Software Engineering

   • Code pattern analysis
   • Best practices extraction
   • Documentation generation
   • Technical support automation

Data Quality

• Completeness: 97.47% (excellent)
• Uniqueness: 99.99% (excellent)
• Answer Coverage: 54.79% (good)
• Code Presence: 92.62% (excellent)
• Overall Quality Score: 53.65/100

License

This dataset is licensed under CC-BY-SA-4.0 (Creative Commons Attribution-ShareAlike 4.0 International), matching Stack Overflow's content license.

Citation

If you use this dataset in your research, please cite:

@dataset{stackoverflow_programming_questions_2025,
 title = {Stack Overflow Programming Questions Dataset (2020-2025)},
 author = {kutayahin},
 year = {2025},
 url = {https://www.kaggle.com/datasets/kutayahin/stackoverflow-programming-questions-2020-2025},
 license = {CC-BY-SA-4.0}
}

Acknowledgments

• Data collected from Stack Overflow via Stack Exchange API
• Stack Overflow community for providing valuable Q&A content
• Stack Exchange for providing public API access

Updates

• Version 1.0 (2025-11-15): Initial release with 95,636 questions from 20 programming languages

Contact

For questions, suggestions, or issues, please open an issue on the dataset page or contact the dataset maintainer.

Related Datasets

• Stack Over...

Reach: Invitations delivered, unique attendees, live vs. on-demand views Quality of participation: Question diversity, participation rate, average watch time Understanding and sentiment: Pre/post comprehension, confidence, and support metrics Issue resolution: Percentage of questions answered, time to closure on action items Trust signals: Follow-up engagement, opt-in rates, advocacy indicators

This statistic represents the importance for obtaining answers on questions of faith among Lebanese Millennials as of 2017. During the survey, ** percent of Lebanese Millennials stated that obtaining an answer for their questions of faith was very important to them.

Aggregate statistics of distributions across the three sites of answerers answering different number of unique questions.