Tibu.ai - Rule Based Telegram Chatbot

Microsoft/ Africa Leadership Academy / mDoc Health-Tech Hackathon.

Background

Many patients in Kenya find it difficult to locate a qualified healthcare specialist around their area. This forces them to travel long distances to find one, possibly suggested to them by a friend, another physician or found online. This requires more time and usually costs more. Also, it is difficult to ascertain the credibility of the said physician.

Solution:

Create a Chatbot that allows users to receive suggestions on recommended healthcare facilities or providers based on their symptoms, diagnosis or test results. The suggestions will be generated from available datasets and a stored database of frequently asked questions (FAQs) from previous user interactions with the chatbot.

Chatbot Functionality

Implementation has been divided into 3 stages. The bot should be able to:

Stage 1:

1. Recommend a facility/provider based on symptoms.

2. Recommending a facility/provider based on specialty (as a redundancy)

Stage 2

1. Recommending a facility/provider based on location.

Stage 3

1. Recommending a facility/provider based on insurance type.

2. Recommending a facility/provider based on operation hours and days.

Datasets

Included in this projects are the following datasets:

specialist.xlsx

This dataset has over 13,000 health facilities and contains:

1. Name of facility
2. Level of care
3. Facility type
4. Ownership
5. Opening days
6. Opening times
7. Geodata
8. Number of beds
9. Insurance providers accepted

symptoms.csv

This Kaggle dataset by Pranay Patil has columns containing diseases, their symptoms , precautions to be taken, and their weights.

diag.spec.icd10.csv

This is a modified ICD-10 dataset where the healthcare experts in our group matched diagnoses with specialists.

Roadmap

In future, the bot should transition from a rule based chatbot into one that utilizes Machine Learning.

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

Authors and Acknowledgement

Great thanks to each of the Group 13 Members for their work and effort in the project:

1. Monica Oyugi
2. Bruce Onduru
3. Hezron Munyakin
4. Fred Mutisya
5. Wayne Asava

License

GNU GPLv3