← All proposals
Illustration: Researchers tracing an automated moderation decision to a human appeal

From moderation decision to appeal

Check whether platform moderation records explain a decision well enough to support scrutiny and appeal.

Download brief

Technical

Bachelor’s · applied project

Final result

Working ETL, automated validation suite, source-linked dashboard and defect list.

02

What to do

  • Build a data-quality dashboard for official DSA moderation records.
03

How it is tested

Use 20,000 records from two platforms over the same 30-day period. Run schema, missingness, timing and cross-field checks; manually verify 100 stratified records.

04

Before you start

Example tools · Python, httpx, pandas or Polars, Pandera, DuckDB, Plotly Dash or Streamlit These are suggestions; equivalent tools are fine.

Access or people · Confirm API limits and current schema; one independent reviewer for the 100-record check.

Research method · Build and test · 6 steps
  1. Read the starting sources and choose one established implementation method.
  2. Write the requirements, data fields, system diagram and test cases.
  3. Prepare 20,000 DSA statements of reasons. Make the answer key and pass criteria before testing.
  4. Build a working version of working etl, automated validation suite, source-linked dashboard and defect list.
  5. Run function, integration and failure-case tests. Record each result.
  6. Run the practical evaluation and list the changes the system still needs.
Download the complete bachelor planDownload bachelor instructions

Master’s · research project

Final result

Versioned ETL, validation rules, source-linked audit dashboard and data-quality report.

01

Dataset

100,000 DSA statements of reasons from four platforms

Use the official Research API to sample 25,000 records per named platform from the same fixed 60-day period. Store record UUID, platform, decision dates, restriction, content type, legal/terms ground, category, automation flags and source permalink. The database contains reporting metadata only.

02

What to do

  • Build a versioned Research API ETL with schema, enumeration, cross-field and provenance checks over a fixed four-platform sample.
  • Validate rule outputs against a stratified human audit, preserve record-level traceability and distinguish formal validity from practical contestability.
  • Report missingness, invalid combinations, category entropy, delay and automation-use differences with uncertainty; never treat self-reported records as ground truth.
03

How it is tested

Pin API query, schema and retrieval date. Validate completeness and allowed combinations; hand-check a stratified sample of 400 records. Compare rule-only quality scores with a simple anomaly detector as a secondary method. Report missingness, invalid-combination rate, category entropy, reporting delay and automation-use differences with platform-level confidence intervals. Treat each record as a platform report and verify the fields that can be checked against the published schema.

04

Before you start

Example tools · Python, httpx, pandas or Polars, Pandera, DuckDB, Plotly Dash or Streamlit These are suggestions; equivalent tools are fine.

Access or people · Confirm current API limits and data-retention window; independent review of 400 sampled records.

01—03

Start with these three actions

  1. Read the API and schema; save one small response with retrieval time and query before defining checks.
  2. Name four platforms and one 60-day window; finalise fields, allowed values, exclusions and the 400-record audit sample.
  3. Implement schema and cross-field rules before aggregation; make every chart trace back to record UUIDs and source permalinks.
Files, questions and sources

Pilot question: Which reporting choices make automated moderation decisions observable and contestable across platforms?

Literature search: DSA transparency database statement of reasons data quality automated moderation audit appeal

Download starter instructions
Research method · Design, build and test (DSR) · 7 steps
  1. Read the newest papers and list the closest existing systems.
  2. Write down the versions, fields, data split, case assignment, random seeds and correct answers for 100,000 DSA statements of reasons from four platforms.
  3. Draw the user workflow, data model and system architecture. List the requirements and pass criteria.
  4. Build a working version of versioned etl, validation rules, source-linked audit dashboard and data-quality report.
  5. Test every function, connection and failure case. Save the failed tests as well as the passed tests.
  6. Compare the system with the named alternative. Then run the user task or decision task in the assignment.
  7. Report the measured result, the failed cases and the design lessons another team can reuse.
Research question and sources

Research question: Design and evaluate an audit method that distinguishes formal schema compliance from records that actually support scrutiny and appeal.

Research result: A validated audit method, cross-platform evidence on reporting quality and design requirements for contestable moderation records.

Current project literature

IS theory starting points

Search Scopus or Web of Science and ACM Digital Library using the topic query, then follow citations to the thesis start date. Record searches and compare methods, data, findings and limitations in literature-matrix.csv. Use that review to confirm or revise the gap and choose a current comparator. The linked papers are starting points.

Search terms: DSA transparency database statement of reasons data quality automated moderation audit appeal

Non-technical

Bachelor’s · applied project

Final result

Coded dataset, comparison report and improved explanation template.

02

What to do

  • Check whether official moderation records explain what happened, why it happened and how the decision can be challenged.
03

How it is tested

Code 120 DSA records from two platforms with a fixed checklist. Double-code 24 records and compare missing information by platform and decision type.

04

Before you start

Example tools · LibreOffice Calc, Zotero, Taguette or NVivo; no programming required. These are suggestions; equivalent tools are fine.

Access or people · A second coder for 24 records; no participant recruitment.

Research method · Study and improve · 6 steps
  1. Read the starting sources and write the exact information problem.
  2. Prepare 120 DSA statements of reasons + six policy pages and a separate answer sheet or coding sheet.
  3. Run one pilot and fix unclear questions.
  4. Collect the named evidence with consent.
  5. Group the findings with the stated categories and check the answer sheet.
  6. Produce coded dataset, comparison report and improved explanation template. List the three most useful changes.
Download the complete bachelor planDownload bachelor instructions

Master’s · research project

Final result

A scored record table, an appeal-process diagram and a template for explanations and appeal information.

01

Dataset + material

240 DSA records + 12 platform policy documents + 8–10 professional walkthroughs

Select 60 records per platform from one fixed period, balanced by decision ground and automation field. Add each platform’s terms, moderation policy and appeal page. Run fictional walkthrough cases with professionals and material free of harmful content.

02

What to do

  • Select 240 DSA statements of reasons across four platforms, six decision grounds and automated/manual decisions.
  • Code whether each record states the action, rule, factual basis, automation role and a usable path to challenge the decision.
  • Compare the records with platform appeal policies and run 8–10 professional walkthroughs of fixed fictional cases.
03

How it is tested

Apply a fixed coding guide to all records and double-code 20%. Compare completeness and contestability by platform, ground and automation field. In walkthroughs, ask professionals to explain the decision and choose an appeal step; code missing information, responsibility and due-process concerns.

04

Before you start

Example tools · LibreOffice Calc, Zotero, Taguette or NVivo; no programming required. These are suggestions; equivalent tools are fine.

Access or people · Recruit 8–10 researchers, regulators, civil-society or prevention professionals; obtain consent and avoid confidential cases.

01—03

Start with these three actions

  1. finalise the four platforms, period, sampling table and 12 policy documents; archive the source files.
  2. Pilot the coding guide on 24 records with two coders; revise and finalise the guide before the final sample.
  3. Write three fictional walkthrough cases and one fixed appeal task; recruit professionals only after consent and data handling are approved.
Files, questions and sources

Pilot question: When does a platform’s explanation give an outsider enough information to understand and challenge a moderation decision?

Literature search: content moderation appeal algorithmic justice DSA statements of reasons contestability platform governance

Interview prompts

  • Can you explain what happened and why from this record alone?
  • What information would you need before advising an appeal?
  • Who is responsible for checking or correcting this decision?
Download starter instructions
Research method · Study, compare and explain · 7 steps
  1. Read the newest papers and write one exact research question.
  2. State which people, cases or documents you will study and what you will compare.
  3. Write the selection rules, questions and analysis steps for 240 DSA records + 12 platform policy documents + 8–10 professional walkthroughs.
  4. Run one pilot. Fix unclear questions or categories, then keep the guide unchanged.
  5. Collect the named interviews, cases or documents with consent.
  6. Analyse them with the stated comparison or coding method. Keep disagreements and missing data.
  7. Report the answer, the evidence and the practical output named in the assignment.
Research question and sources

Research question: Explain how automated moderation records distribute responsibility and due process between platforms, reviewers, regulators and users.

Research result: A theory-linked model of contestable moderation and evidence-based requirements for explanations and appeals.

Current project literature

IS theory starting points

Search Scopus or Web of Science and ACM Digital Library using the topic query, then follow citations to the thesis start date. Record searches and compare methods, data, findings and limitations in literature-matrix.csv. Use that review to confirm or revise the gap and choose a current comparator. The linked papers are starting points.

Search terms: content moderation appeal algorithmic justice DSA statements of reasons contestability platform governance

Your degree, track and shortlist are saved in this browser.