Created with Gpt o4-mini-high

ChatGPT’s new model has changed different industries. But I think this new model will shift Data Science the most, why?

Because it would automate the coding process, this change will go beyond any changes made.

In this article, I will first split Data Science into multiple disciplines, starting with Data Exploration and going to Machine Learning. But before doing them, let’s see the data we will use.

Bitcoin Price Analysis

Now let’s use this dataset from Kaggle.

Reference

Good, now download the data

Now let’s start with Data Exploration. Click on the “Download” button on the top right.

Download the dataset

Good, now you’ll have those files.

Datasets

I selected Binance.csv and uploaded it to the GPT-o4-mini-high.

o4-mini-high

But which prompt should we use? I would use Prompt Perfector from our platform to craft a perfect prompt for data exploration.

Prompt Perfector

Reference

Here is the link to use this assistant. I attached the dataset to the prompt perfector + key features from here, with this prompt:

I want to develop a prompt for data exploration

Here is the output.

Prompt Perfector

As you can see, we can make it perfect by following the perfection steps. But this should be enough, so let’s use this prompt.

Data Exploration

Photo by Maël BALLAND on Unsplash

Here we’ve created this prompt.

Explore the BTCUSD combined index dataset, 
which contains 1-hour historical data from multiple exchanges, 
including open, high, low, close prices, volume, and trade information. 
Provide an initial data exploration, including summary statistics, 
missing value checks, and visualizations of key features such as price trends 
and trading volume over time.

Let’s see the process.

SS of the output

But the process did not end.

SS of the output

I like it, and the best part is, it only lasted 12 seconds. Can you believe it? In 12 seconds, ChatGPT’s new model automates Data Exploration and says hello to us, Data Scientist.

Let’s continue with the Data Visualization.

Data Visualization

Photo by Clay Banks on Unsplash

Now I repeated the process with the Prompt Perfector to craft a data visualization prompt. Here is the prompt it gave us.

Using the BTCUSD combined index dataset (with columns like Open time, Close, Volume, and Number of trades), create three resource-efficient visualizations:

Daily Closing Price Line Chart: Aggregate the data to daily frequency and plot the closing price to show overall price trends.
Daily Total Trading Volume Bar Chart: Aggregate trading volume by day and plot as a bar chart to highlight periods of high or low activity.
Scatter Plot of Daily Number of Trades vs. Daily Volume: For each day, plot the total number of trades against total trading volume to explore their relationship.

Let’s explore the output more in detail.

Here is the first part- thinking.

SS of the output

Here is the second part- action.

Graphs

Machine Learning

Photo by Markus Winkler on Unsplash

Now let’s craft the machine learning prompt for the machine learning with the prompt perfector.

Here is the prompt.

Using the BTCUSD combined index dataset (with columns such as Open time, Close, and others), build a machine learning model to predict the closing price for the next 7 days. The model should be trained on historical data, and the output should include:

A line graph showing both the actual and predicted closing prices for the upcoming week, plotted on the same chart for easy comparison.
Efficient data handling to accommodate the large dataset (e.g., downsampling to daily data, using only relevant columns).

Now let’s see the output.

SS of the output

And in 10 seconds.

LearnAIWithMe Agent

Let’s upload it to our agent from here.

Now, here you can explore the data, visualize it, and ask questions by using our AI assistant.

Or build a machine learning model by selecting a model name or other variables, like this;

SS of the otuput.

Final Thoughts

And we all did in 28 seconds? From Data exploration to machine learning, the o4 mini-high applies and evaluates all codes in just 28 seconds. Here are the steps;

• Data Exploration — 12 seconds
• Data Visualization — 6 seconds
• Machine Learning — 10 seconds

One word: amazing. I amazed. Apart from other models, it also thinks before applying codes.

To follow similar articles, use our assistants or automates, visit our platform.

Here are the free resources.

Here is the ChatGPT cheat sheet.

Here is the Prompt Techniques cheat sheet.

Here is my NumPy cheat sheet.

Here is the source code of the “How to be a Billionaire” data project.

Here is the source code of the “Classification Task with 6 Different Algorithms using Python” data project.

Here is the source code of the “Decision Tree in Energy Efficiency Analysis” data project.

Here is the source code of the “DataDrivenInvestor 2022 Articles Analysis” data project.

“Machine learning is the last invention that humanity will ever need to make.” Nick Bostrom