How I Used Ai To Analyse 40 Years Of Rainfall In Dormaa East In Less Than 30 Minutes

In less than 30 minutes, I pulled and made a visual of all the days it rained in Dormaa East since I was born (1985) up to now (2025). That’s a 40-year span.

GitHub Copilot wrote the entire python code in a split second. All I did was to pip install the requests library. I didn’t even bother going to look for API endpoints and what not. It automatically connected to NASA’s API using the right coordinates for Dormaa East and dumped the results on my desktop as a CSV file.

Below is the prompt

The only thing I didn’t like about the csv data was the date format. Instead of 01/01/1985, it gave me 19850101. I asked it to correct it, but it produced the same thing again. I didn’t push further (to avoid hallucination), so I solved it in Excel.

And what I did was simple: I split the 19850101 value into day, month, and year using RIGHT, MID, and LEFT, then I again used the DATE formula to reassemble it:

Something like this:

You may be wondering, how did this thought even occur to me?

The reason is the Murabit Project (Murabit.org)

We are exploring the most economic way to get the best outcomes for the children. One thing we realised quickly is that if you’re not careful with projects like this, most of the funds go into structures, not children’s learning. The Brone Krom structure in Wamfie for instance costs us GHS 25,000. Imagine how many centers that will build if we can setup one at a cost of GHS 5,000.

One thing we considering now is Mobile Library. The idea is simple:

Use plywood to make a fenced reading area.
Put in books, chairs, a teacher, and a whiteboard.
Motivate the children to come and read.
After about 3 months, move it to another community.

By the end of the year, one setup can cover 4 communities. If we have 3 setups, that’s 12 communities per year. If each child studies 3 hours per day, that is 1,095hous. If you get 30 children, that will be 32,850hrs of learning.

But that is the ideal scenario. The reality is far different.

The first question is: What if it rains?
If we go by the design we are imagining, it means anytime it starts to rain, we would have to close the children.

This is what led me to check how many times it has rained in Dormaa East since I was born. With that data, we’ll be able to estimate:

How many days per year are rain-free.
The average learning hours children can realistically achieve.

Now, ;let's come back to the data.
When I opened the Excel/CSV file, I asked AI again to classify the rainfall values for decision-making. At first, it gave me:

Drizzling
Steady Rain
Heavy Rain

But when I asked for globally recognised groupings, it gave me:

No Rain
Light Rain
Moderate Rain
Heavy Rain
Very Heavy Rain
Extreme Rainfall

"These thresholds are used by agencies like the World Meteorological Organization (WMO) and national weather services."- AI said.

I then asked it to use a SWITCH case in Excel, with rainfall values in column B. It produced this:

I pasted it in column B, and it worked like Mohammed Kudus!

With the transformed data now, I did the visual and the pivot table below:

The final thing I did was to take a screenshot of the visual and paste in AI and asked it to analyze.

I asked further questions but unfortunately it told me:

All this that I've talked about took me less than 30min.

Just 3 years ago, this same task would have taken me a day or two. That means I have saved about 99.99% of time.

This is why I always say: After the invention of the Internet, AI is the most valuable invention for humanity.

See you again :)