aie_recsys_cap

AIEdge Recommender System Capstone (June 2023 Cohort)

1. create a virtual env in the /streamlit folder

python -m venv .venv

OR

conda create -n tf_recsys

2. activate the virtual env

source .venv/bin/activate

OR

conda activate tf_recsys

3. In the streamlit folder, pip install streamlit version 1.22, tensorflow and tensorflow-recommenders

pip install streamlit==1.22
pip install tensorflow==2.11.1
pip install tensorflow-recommenders

4. Test streamlit installation

streamlit hello

Quit the streamlit application if it runs OK.

5. Run the streamlit app file app.py is input user_ID, output 10 game recs

app_e.py is input list of game appids, output 10 game recs

streamlit run app.py
streamlit run app_e.py

Source of Datasets

(final cleaned dataset was a lowercase concatenated game name join between steam.csv and steam-200k.csv)

1. Nik Davis Dataset of 27,000 Steam games and their metadata

https://www.kaggle.com/datasets/nikdavis/steam-store-games

2. Tamber Dataset of 200,000 Steam user interactions (Play, Purchase)

https://www.kaggle.com/datasets/tamber/steam-video-games

requirements.txt

altair==4.2.2
protobuf==3.19.6
python==3.10.12
scann==1.2.9
streamlit==1.22.0
tensorflow==2.11.1
tensorflow-recommenders==0.7.3

for running the notebooks

jupyter==1.0.0