Skip to content
View CarolineKarimi's full-sized avatar

Achievements

Achievement: YOLOAchievement: Pull Shark

Achievements

Achievement: YOLOAchievement: Pull Shark

Block or report CarolineKarimi

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
CarolineKarimi/README.md

Hi, I'm Caroline 👋

I turn messy data into decisions someone can actually act on — credit risk models that work with the data a lender actually has, and charts that don't need explaining.

Currently:

  • Building out Phases 3 and 4 of the credit risk lifecycle — real-time data infrastructure and deployment
  • Exploring how digital lenders can clearly communicate loan obligations on uncollateralized credit
  • Using AI and Streamlit to build dynamic, decision-ready dashboards
  • Still, always, happiest turning a cluttered chart into an obvious one

What I've built

credit-risk-lifecycle — the full arc of a lending product's risk engine: a cold-start expert scorecard, then a data-driven model trained on real repayment outcomes (Random Forest, 80% accuracy), phase by phase.

Storytelling-with-Data — eight charts, rebuilt twice each: once the way most tools default to, once the way they should look when the point actually lands.

Also here: SQL fundamentals, clustering techniques, and fraud/anomaly detection — practical Python notebooks I keep as working reference.

Toolbox

Python (pandas, scikit-learn) · SQL · Streamlit · Chart.js · credit risk modeling · M-Pesa/telco & bureau data · Kenya fintech & digital lending

LinkedIn · Email

Pinned Loading

  1. Storytelling-with-Data Storytelling-with-Data Public

    Rebuilding cluttered charts as clear ones — chart pairs demonstrating showing data vs storytelling-with-data.

    HTML

  2. credit-risk-lifecycle credit-risk-lifecycle Public

    A lending product's risk engine, phase by phase — from a cold-start expert scorecard to a data-driven model trained on real repayment outcomes.

    Jupyter Notebook

  3. Data_Anomalies_and_Fraud_Detection_Python Data_Anomalies_and_Fraud_Detection_Python Public

    Flagging fraudulent or anomalous data with four detection methods — 3-sigma rule, isolation forest, Benford's Law, and correlation heat maps — applied to catch suspect records.

    Jupyter Notebook

  4. clustering_in_python clustering_in_python Public

    Comparing clustering techniques — k-means, hierarchical, DBSCAN, and more — for grouping and segmenting data.

    Jupyter Notebook

  5. SQL_Tutorial-Run_in_Jupyter_Notebook SQL_Tutorial-Run_in_Jupyter_Notebook Public

    A hands-on SQL tutorial run entirely inside Jupyter Notebook — building a database with SQLAlchemy and querying it with real SQL commands.

    Jupyter Notebook