\n\n\n\n
![\"\"](\"https:\/\/tbekk.com\/devstream\/wp-content\/uploads\/2023\/07\/image-1024x576.png\")
<\/figure>\n\n\n\nAnalyzing a dataset to understand it better and gain insights is an important step in any data science project. And ChatGPT can help simplify and speed up such analysis.<\/p>\n\n\n\n
Here we\u2019ll perform exploratory data analysis on a sample dataset using ChatGPT. We\u2019ll understand distributions of features, how a feature of interest varies with other features, and much more. <\/p>\n\n\n\n
To follow along and access ChatGPT, you need an OpenAI account. If you don’t have one, you can sign up for a free account<\/a>.<\/p>\n\n\n\n The seaborn library<\/a> has some ready-to-use datasets that we can analyze. We\u2019ll use the tips<\/strong> dataset. <\/p>\n\n\n\n Prompt<\/strong>: Tell me more about the `tips` dataset built into the seaborn library. Knowing what the dataset is about is a good starting point. Let’s start by doing the following:<\/p>\n\n\n\n Prompt<\/strong>: Let\u2019s use the `tips` dataset. Write the code to read the dataset into a pandas dataframe and get the basic info and summary statistics of the available fields. We see that we have quite an informative summary of the dataset:<\/p>\n\n\n\n From the summary statistics, we have an idea of the numerical features in the dataset. We know the minimum and maximum values, mean and median values, and percentile values for the numerical features. There are no missing values so we can proceed with the next steps.<\/p>\n\n\n\n Now that we have an idea of the dataset, let’s go further.<\/p>\n\n\n\n The goal of this exploratory data analysis is to understand the tipping behavior better. For this we can come up with helpful visualizations. These should help us understand the relationship of the tip amount to the various categorical variables in the dataset.<\/p>\n\n\n\n Because this is a simple dataset to analyze, let’s prompt ChatGPT to give us a set of steps to go about analyzing this data set further.<\/p>\n\n\n\n Prompt<\/strong>: The goal of the analysis is to get some insights into the tipping behavior, starting with some simple visualizations. Suggest steps for the same. The data exploration steps suggested by ChatGPT all seem valid. So we will use these steps\u2014to better understand a dataset\u2014one step at a time. We\u2019ll prompt ChatGPT to generate code, try to run the generated code, and modify them as needed.<\/p>\n\n\n\n As a first step, let’s visualize the distribution of the tip amount prompt. <\/p>\n\n\n\n Prompt<\/strong>: Write the code to plot the distribution of tip amounts. ChatGPT gave the following code to generate a histogram and a kernel density plot that help us get an idea of the distribution of the tip amounts:<\/p>\n\n\n\n We have many categorical variables in the dataset. It will help to have a simple bar plot showing the average tip amount for the different values of these category variables.<\/p>\n\n\n\n Prompt<\/strong>: Use all the available categorical variables (\u2018sex\u2019, \u2018smoker\u2019, \u2018day\u2019, and \u2018time\u2019). Write the code to generate bar plots to understand tipping behavior. Use subplots so we can have all the four bar plots in a single figure. And here\u2019s the code. The code ran without errors; I only had to modify the figure size:<\/p>\n\n\n\n From the plots, we see that features like sex and smoking behavior don\u2019t influence tipping behavior (which is expected). While days and times seem to. The average tip amount on weekends and dinner seem to be slightly higher.<\/p>\n\n\n\n Now, let\u2019s see how the total bill influences the tip amount paid.<\/p>\n\n\n\n Prompt<\/strong>: I\u2019d like to understand the relationship between total bill and the tip amount. Please give me the code to generate a suitable plot for this. I believe a simple scatter plot will be helpful. Here\u2019s the code to generate the required scatter plot:<\/p>\n\n\n\n We see that there\u2019s some positive correlation between the total bill and the tip amounts.<\/p>\n\n\n\n As a next step, let\u2019s try visualizing the relationship between the party size and the tip amount. <\/p>\n\n\n\n Prompt<\/strong>: Now, I’d like to understand how the tip amount varies with the number of people who dine (party size). Which is a better choice: a bar plot or a violin plot? As we\u2019d also like to understand the distribution of the tip amounts, let\u2019s go ahead with creating a violin plot.<\/p>\n\n\n\n Prompt<\/strong>: Cool! Please write the code to generate a violin plot for this visualization.<\/p>\n\n\n\n And here\u2019s the code:<\/p>\n\n\n\n Next, let\u2019s understand how the time and<\/em> the day influence the tipping behavior.<\/p>\n\n\n\n Prompt<\/strong>: I want to understand how the tipping behavior varies depending on both time and day. What is a good visualization? Also, write the code to generate the required visualization. This code snippet shows how to pivot using the day as the index and aggregate the mean value of the tip amounts for lunch and dinner (the time of the day):<\/p>\n\n\n\n From the heat map, we see that the average tip amounts are marginally higher for dinner on Sunday than other time-day pairs. In general, for a given day, on an average, tip amounts for dinner seem to be higher than those for lunch. <\/p>\n\n\n\n We also see that there is not enough data for lunch on Saturdays and Sundays. Because we only have a total of 244 records, we may not have any record for lunches on weekends. Let\u2019s go ahead and verify this.<\/p>\n\n\n\n Prompt<\/strong>: I want to get the number of parties who ordered lunch on weekends. Write the code to filter the `tips_data` dataframe and get this count. Here\u2019s the code to filter the dataframe and obtain a subset containing the records for lunches on weekends:<\/p>\n\n\n\n And we see that we don\u2019t have any records for the same, so our analysis is correct thus far:<\/p>\n\n\n\n And that\u2019s a wrap! We explored the `tips` dataset and generated some helpful visualizations by prompting ChatGPT. <\/p>\n\n\n\n In this article, we learned how to leverage ChatGPT for data exploration. If you\u2019re interested in integrating ChatGPT into your data science workflow, check out this guide<\/a>. It walks through an example project\u2014along with tips and best practices\u2014to effectively use ChatGPT for data science experiments.Loading the Dataset and Getting Basic Info<\/h1>\n\n\n\n
![\"ChatGPT-Powered](\"https:\/\/www.kdnuggets.com\/wp-content\/uploads\/c_chatgptpowered_data_exploration_unlock_hidden_insights_dataset_3.png\")
<\/p>\n\n\n\n\n
![\"ChatGPT-Powered](\"https:\/\/www.kdnuggets.com\/wp-content\/uploads\/c_chatgptpowered_data_exploration_unlock_hidden_insights_dataset_14.png\")
<\/p>\n\n\n\nimport seaborn as sns\nimport pandas as pd\n\n# Load the 'tips' dataset from Seaborn\ntips_data = sns.load_dataset('tips')\n\n# Create a Pandas DataFrame from the loaded dataset\ntips_df = pd.DataFrame(tips_data)\n\n# Display the first few rows of the DataFrame\nprint(\"First few rows of the 'tips' dataset:\")\nprint(tips_df.head())\n\n# Get basic information about the fields\nprint(\"\\nInformation about the 'tips' dataset:\")\nprint(tips_df.info())\n\n# Get summary statistics of the numeric fields\nprint(\"\\nSummary statistics of the numeric fields:\")\nprint(tips_df.describe())<\/code><\/pre>\n\n\n\nFirst few rows of the 'tips' dataset:\n total_bill tip sex smoker day time size\n0 16.99 1.01 Female No Sun Dinner 2\n1 10.34 1.66 Male No Sun Dinner 3\n2 21.01 3.50 Male No Sun Dinner 3\n3 23.68 3.31 Male No Sun Dinner 2\n4 24.59 3.61 Female No Sun Dinner 4\n\nInformation about the 'tips' dataset:\n\nRangeIndex: 244 entries, 0 to 243\nData columns (total 7 columns):\n # Column Non-Null Count Dtype \n--- ------ -------------- ----- \n 0 total_bill 244 non-null float64 \n 1 tip 244 non-null float64 \n 2 sex 244 non-null category\n 3 smoker 244 non-null category\n 4 day 244 non-null category\n 5 time 244 non-null category\n 6 size 244 non-null int64 \ndtypes: category(4), float64(2), int64(1)\nmemory usage: 7.4 KB\nNone\n\nSummary statistics of the numeric fields:\n total_bill tip size\ncount 244.000000 244.000000 244.000000\nmean 19.785943 2.998279 2.569672\nstd 8.902412 1.383638 0.951100\nmin 3.070000 1.000000 1.000000\n25% 13.347500 2.000000 2.000000\n50% 17.795000 2.900000 2.000000\n75% 24.127500 3.562500 3.000000\nmax 50.810000 10.000000 6.000000<\/code><\/pre>\n\n\n\nExploring the Dataset \u2013 The What, the Why, and the How<\/h1>\n\n\n\n
![\"ChatGPT-Powered](\"https:\/\/www.kdnuggets.com\/wp-content\/uploads\/c_chatgptpowered_data_exploration_unlock_hidden_insights_dataset_1.png\")
<\/p>\n\n\n\nExploring the Distribution of Tip Amounts<\/h1>\n\n\n\n
![\"ChatGPT-Powered](\"https:\/\/www.kdnuggets.com\/wp-content\/uploads\/c_chatgptpowered_data_exploration_unlock_hidden_insights_dataset_10.png\")
<\/p>\n\n\n\nimport matplotlib.pyplot as plt\n\n# Create a histogram of tip amounts\nplt.figure(figsize=(8, 6))\nsns.histplot(data=tips_data, x='tip', kde=True)\nplt.title(\"Distribution of Tip Amounts\")\nplt.xlabel(\"Tip Amount\")\nplt.ylabel(\"Frequency\")\nplt.show()<\/code><\/pre>\n\n\n\n![\"ChatGPT-Powered](\"https:\/\/www.kdnuggets.com\/wp-content\/uploads\/c_chatgptpowered_data_exploration_unlock_hidden_insights_dataset_8.png\")
<\/p>\n\n\n\nAnalyzing Tipping Behavior Based on Categorical Variables<\/h1>\n\n\n\n
![\"ChatGPT-Powered](\"https:\/\/www.kdnuggets.com\/wp-content\/uploads\/c_chatgptpowered_data_exploration_unlock_hidden_insights_dataset_12.png\")
<\/p>\n\n\n\n# Define the categorical variables to analyze\ncategorical_vars = ['sex', 'smoker', 'day', 'time']\n\n# Create subplots for each categorical variable\nfig, axes = plt.subplots(nrows=2, ncols=2, figsize=(8, 5))\nfig.suptitle(\"Tipping Behavior based on Categorical Variables\", fontsize=16)\n\n# Generate bar plots for each categorical variable\nfor ax, var in zip(axes.flatten(), categorical_vars):\n sns.barplot(data=tips_data, x=var, y='tip', ax=ax)\n ax.set_xlabel(var.capitalize())\n ax.set_ylabel(\"Average Tip Amount\")\n\nplt.tight_layout()\nplt.show()<\/code><\/pre>\n\n\n\n![\"ChatGPT-Powered](\"https:\/\/www.kdnuggets.com\/wp-content\/uploads\/c_chatgptpowered_data_exploration_unlock_hidden_insights_dataset_9.png\")
<\/p>\n\n\n\nVisualizing the Relationship Between Total Bill and Tip Amount <\/h1>\n\n\n\n
![\"ChatGPT-Powered](\"https:\/\/www.kdnuggets.com\/wp-content\/uploads\/c_chatgptpowered_data_exploration_unlock_hidden_insights_dataset_5.png\")
<\/p>\n\n\n\n# Create a scatter plot of total bill vs. tip amount\nplt.figure(figsize=(6, 4))\nsns.scatterplot(data=tips_data, x='total_bill', y='tip')\nplt.title(\"Total Bill vs. Tip Amount\")\nplt.xlabel(\"Total Bill\")\nplt.ylabel(\"Tip Amount\")\nplt.show()<\/code><\/pre>\n\n\n\n![\"ChatGPT-Powered](\"https:\/\/www.kdnuggets.com\/wp-content\/uploads\/c_chatgptpowered_data_exploration_unlock_hidden_insights_dataset_11.png\")
<\/p>\n\n\n\nVisualizing the Relationship Between Party Size and Tip Amount <\/h1>\n\n\n\n
![\"ChatGPT-Powered](\"https:\/\/www.kdnuggets.com\/wp-content\/uploads\/c_chatgptpowered_data_exploration_unlock_hidden_insights_dataset_2.png\")
<\/p>\n\n\n\n# Create a violin plot for tip amount by party size\nplt.figure(figsize=(6, 4))\nsns.violinplot(data=tips_data, x='size', y='tip')\nplt.title(\"Tip Amount by Party Size\")\nplt.xlabel(\"Party Size\")\nplt.ylabel(\"Tip Amount\")\nplt.show()<\/code><\/pre>\n\n\n\n![\"ChatGPT-Powered](\"https:\/\/www.kdnuggets.com\/wp-content\/uploads\/c_chatgptpowered_data_exploration_unlock_hidden_insights_dataset_7.jpg\")
<\/p>\n\n\n\nUnderstanding Tipping Behavior Based on Both Time and Day<\/h1>\n\n\n\n
![\"ChatGPT-Powered](\"https:\/\/www.kdnuggets.com\/wp-content\/uploads\/c_chatgptpowered_data_exploration_unlock_hidden_insights_dataset_13.png\")
<\/p>\n\n\n\n# Create a pivot table of average tip amount by time and day\npivot_table = tips_data.pivot_table(values='tip', index='day', columns='time', aggfunc='mean')\n\n# Create a heatmap of tipping behavior based on time and day\nplt.figure(figsize=(8, 6))\nsns.heatmap(pivot_table, cmap='YlGnBu', annot=True, fmt=\".2f\", cbar=True)\nplt.title(\"Tipping Behavior based on Time and Day\")\nplt.xlabel(\"Time\")\nplt.ylabel(\"Day\")\nplt.show()<\/code><\/pre>\n\n\n\n![\"ChatGPT-Powered](\"https:\/\/www.kdnuggets.com\/wp-content\/uploads\/c_chatgptpowered_data_exploration_unlock_hidden_insights_dataset_6.jpg\")
<\/p>\n\n\n\n![\"ChatGPT-Powered](\"https:\/\/www.kdnuggets.com\/wp-content\/uploads\/c_chatgptpowered_data_exploration_unlock_hidden_insights_dataset_15.png\")
<\/p>\n\n\n\n# Filter the data for lunch on weekends\nweekend_lunch_data = tips_data[(tips_data['time'] == 'Lunch') & (tips_data['day'].isin(['Sat', 'Sun']))]\n\n# Get the count of parties who ordered lunch on weekends\ncount = weekend_lunch_data['size'].count()\nprint(\"Number of parties who ordered lunch on weekends:\", count)<\/code><\/pre>\n\n\n\nNumber of parties who ordered lunch on weekends: 0<\/code><\/pre>\n\n\n\nWrapping Up<\/h1>\n\n\n\n
Bala Priya C<\/a><\/strong> is a developer and technical writer from India. She likes working at the intersection of math, programming, data science, and content creation. Her areas of interest and expertise include DevOps, data science, and natural language processing. She enjoys reading, writing, coding, and coffee! Currently, she’s working on learning and sharing her knowledge with the developer community by authoring tutorials, how-to guides, opinion pieces, and more.<\/p>\n","protected":false},"excerpt":{"rendered":"