Why this matters
Once your analysis is useful, people will ask the same question again with a different district, month or product. Rather than rerunning code for every request, you can give them a dashboard or a small data app to explore on their own. Adverts name dashboards often, with Power BI and Tableau common and Shiny or similar app tools for some roles. In Ghana there is an extra test: will it load on a phone over mobile data in Bolgatanga?
Choose the lightest tool that works
- A one page summary or PDF: best when the decision is made once.
- A spreadsheet with a pivot table: fine for small data and familiar to most managers.
- A dashboard tool such as Looker Studio, Power BI or Tableau: good for regular reporting with filters. You met these in the Data Analyst track.
- A small data app written in Python: useful when people need to run a model or a calculation, such as "what if we lend to 50 more farmers?"
Streamlit is one example of a free, open source Python library for data apps. A minimal app looks like this:
import pandas as pd
import streamlit as st
@st.cache_data
def load():
return pd.read_csv("district_summary.csv")
df = load()
st.title("Repayment by district")
district = st.selectbox("District", sorted(df["district"].unique()))
st.bar_chart(df[df["district"] == district], x="month", y="on_time_rate")