Event datasets used for training machine learning models.
For each supported language, learning events and assessment events are continuously uploaded from Android devices to the webapp database.
And those datasets are then downloaded to this repository in a daily cron job.
flowchart
    subgraph device [Android device]
        kukariri -- AssessmentEvents --> analytics@{ shape: cyl }
    
        subgraph sg_lit [Literacy apps]
        herufi
        vitabu
        etc_lit[...]
        end
        sg_lit -- LearningEvents --> analytics
    
        subgraph sg_num [Numeracy apps]
        nyas-space-quest
        nambari
        etc_num[...]
        end
        sg_num -- LearningEvents --> analytics
    end
    subgraph rest_api_hin [hin.elimu.ai]
        analytics-->webapp_hin[webapp]@{ shape: cyl }
    end
    webapp_hin-->ml-datasets
    subgraph rest_api_tgl [tgl.elimu.ai]
        webapp_tgl[webapp]@{ shape: cyl }
    end
    webapp_tgl-->ml-datasets
    subgraph rest_api_tha [tha.elimu.ai]
        webapp_tha[webapp]@{ shape: cyl }
    end
    webapp_tha-->ml-datasets
    subgraph rest_api_vie [vie.elimu.ai]
        webapp_vie[webapp]@{ shape: cyl }
    end
    webapp_vie-->ml-datasets
    click kukariri "https://github.com/elimu-ai/kukariri"
    click herufi "https://github.com/elimu-ai/herufi"
    click vitabu "https://github.com/elimu-ai/vitabu"
    click nyas-space-quest "https://github.com/elimu-ai/nyas-space-quest"
    click nambari "https://github.com/elimu-ai/nambari"
    click analytics "https://github.com/elimu-ai/analytics"
    click webapp_hin "https://github.com/elimu-ai/webapp"
    click webapp_tgl "https://github.com/elimu-ai/webapp"
    click webapp_tha "https://github.com/elimu-ai/webapp"
    click webapp_vie "https://github.com/elimu-ai/webapp"
    When machine learning models are being trained with datasets collected from the elimu.ai Android apps, they should be fetching the data from this repository.
Note
This repository only contains learning events and assessment events. If you are also looking for content datasets, see webapp-lfs .
You can expect the datasets in this repository to be updated once per day.
Tip
Since datasets in this repository are continuously updated, you should also design your machine learning code to continulously train new versions of your model (e.g. once per night).
Prerequisites:
- Install Python
 
Install the Python dependencies:
pip install -r requirements.txtDownload datasets:
python download_datasets.pyelimu.ai - Free open-source learning software for out-of-school children πβ¨
Website π Β β’Β Wiki π Β β’Β Projects π©π½βπ» Β β’Β Milestones π― Β β’Β Community ππ½ Β β’Β Support π
