NetworkScience

This Git repository was created for the final project.

Data can be consumed by this Streamlit app. Source code available at Github.

Steps to analyze a network using Neo4j

1. Go to Neo4j Labs
2. Drop file "Neo4j-a29b25df-Created-2024-11-15.txt" to use as credentials. This will connect to an AuraDB
3. Download previous tests at Inep Website
4. Drop it at Neo4j Labs on "Drag & Drop"
5. Click on "Graph Enhancement"
6. On Node Labels use: subjects, study topics, behaviours. Hint: type each one separatadely and then press "tab"
7. On Relationship Types use: related to, discussed about, talked about, includes, needs, contains. Use the same hint as previous point.
8. Select document and then click on "Generate Graph"
9. Select document again and click on "Explore graph" this will open Neo4j Bloom
10. Go to "Query" tab: if you click on any "Database information" item, this will create a Cypher query that will run in the right panel.
11. In the right panel there is an option called "Table" if you go there, there is a small "download" icon, that allows you to download the graph as a *.json file
12. Use the jupyter notebook to process this data

Another data sources:

Maritaca Enem Enem Microdata

General Notes

Maritaca dataset has only data from those two specific tests:

• ENEM_2022_P1_CAD_03_DIA_1_BRANCO

  • 1º Dia – Caderno 3 – Branco - Aplicação Regular
    • Prova de linguagens, códigos e suas tecnologias e redação (Test code: 1068)
    • Prova de ciências humanas e suas tecnologias (Test code: 1057)

• ENEM_2022_P1_CAD_06_DIA_2_CINZA

  • 2º Dia – Caderno 6 – Cinza (Superampliada)
    • Prova de ciências da natureza e suas tecnologias (Test code: 1087)
    • Prova de matemática e suas tecnologias (Test code: 1078)

To do:

1. tools/script.py generates treated_data/questions_2022.csv and treated_data/questions_2023.csv [Done]
2. tools/samplig.py generates enem_random_sample.csv and we need to filter tests that matches with Maritaca dataset [Done]
3. tools/dif.py generates enem_dif_dataset.csv [Done]
4. tools/dif.py generates student responses per test in separated *.csv files [Done]
5. Add to all 4 student responses a line with the correct answer [Done]
6. Use the streamlit app to generate statistics of those students [Done]