Request2D aims to help medical researchers to find data files in the dataset or database those are related with their research questions.
For every research project with secondary use of clinical data, researchers concern about the availability of the data in their database, and the location of data columns which are related with the resaerch question to query. For a clinician without data science knowledge, this could be a tough process, traditionally they need help from informationists. ReQuest2D will help them to finish this process all by themselves.
Why ReQuest2D is limited to medicine field? Making use of existing medical ontologies and standards, such as ICD 10, SNOMED, is the key. For example, if we care about glaucoma, the system will search ICD 10 and SNOMED ontology, to find all concepts related with glaucoma, and then go to the database, find all columns that include concepts of our research interest.
What it does
Automatically search related medical concepts for researchers. Map the concepts to datasets. Come with part of data quality service.
How I built it
Step 1. Try find diabetes in ICD 10 - Done Step 2. Manually find those files about diabetes in the database - Done Step 3. See if there is any relationship between the concepts and the data files - Done Step 4. Python to search terminology concepts - Done Step 5. Python to search files -Done The algorithm need to be optimized, tooooooooo slow Ideas: Step 5.1. For big data file: sample to a smaller size - Done Step 5.2. Check column names - Done Step 6. Django project - Done Step 7. GUI framework - Done Step 8. Logo and Homepage - Done Step 9. Make it online available - Done
Challenges I ran into
Django does not support ajax query to do file reading, which was a time killer for me to write
Accomplishments that I'm proud of
This project really solves the most annoying problem that I'm facing with every day. And me alone is a team.
What I learned
I need friends to work together, or it will be too tired.
What's next for ReQuest2D
Finish all functions