4. Data Preparation and Basic Dashboard Construction

This chapter implements dashboards using Claude Code. The forked repository includes sample CSV data files, which we'll use to give specific instructions to Claude Code.

Implementing Dashboards with Claude Code

Let's add charts to the Home.py table. We'll use these two files for analysis:

orders.csv: Order data
users.csv: User data

Establishing Analysis Strategy from Data Content

As the first step, we'll have Claude Code establish an analysis strategy from these files. Give Claude Code instructions like this:

I want to add analysis functionality to @Home.py. Please load the data from @sample_data/users.csv and @sample_data/orders.csv and establish an analysis strategy.

Claude Code will understand the data structure and propose analysis ideas like the following (actual responses may differ slightly from this example):

1. Basic Statistics and KPIs
    - Total users, orders, cancellation rate
    - Age and gender distribution
    - Geographic distribution (by country/state)
    - Traffic source analysis
2. Order Analysis
    - Analysis by order status
    - Monthly and daily order trends
    - Monthly order progression and cancellation rate analysis
3. User Analysis
    - User segmentation (age, gender, region)
    - New users vs. repeat customers
    - Geographic distribution visualization
4. Business Insights
    - Customer lifetime value (estimated)
    - Seasonality analysis
    - Conversion rates by traffic source

If Claude Code's proposals don't align with your intended direction, engage in dialogue at this point to adjust the direction.

Developing Implementation Strategy for Ideas

Tell Claude Code which of the proposed ideas you want implemented. For example, if you choose "monthly order progression and cancellation rate analysis," give instructions like:

Selected idea: Monthly order progression and cancellation rate analysis

Please develop a detailed implementation strategy for this analysis:
- Required libraries
- Data preprocessing methods
- Visualization methods
- Dashboard structure
- Error handling

Claude Code will propose an implementation strategy like:

**Required Libraries**:
- `pandas`: Data manipulation and analysis
- `plotly`: Interactive visualization
- `streamlit`: Dashboard UI
- `openpyxl`: Excel file support

**Data Preprocessing**:
- Date data type conversion
- Creating period data for monthly aggregation
- Handling missing values

**Visualization Methods**:
- Monthly order count: Bar chart
- Monthly cancellation rate: Line chart
- Data table: Interactive display

Development Based on Implementation Strategy

Give the implementation strategy to Claude Code and have it generate actual code (or if Claude Code has already suggested implementing, select Yes and proceed):

Please add the following features:
1. Monthly order count bar chart
2. Monthly cancellation rate line chart

Testing Generated Code

Run the code generated by Claude Code to verify it works:

streamlit run Home.py

If it doesn't work as expected, communicate specific problems and request fixes:

"Date parsing errors are occurring in monthly aggregation"
"Graph axis labels are hard to read"
"Cancellation rate calculation seems incorrect"

By giving specific and clear instructions like this, Claude Code will respond appropriately. If implemented correctly, graphs like this should be displayed: Home.py

Explanation of Libraries Used

Let's cover libraries commonly used in dashboard implementation. The code generated locally probably uses the libraries we'll introduce.

Pandas: Essential for Data Analysis

Pandas is one of the most important libraries in Streamlit BI dashboard development. As the de facto standard library for data analysis in Python, it efficiently handles data loading, processing, transformation, and aggregation.

Pandas Features

1. Loading Various Data Formats

Supports wide range of data formats: CSV, Excel, JSON, SQL, Parquet
Automatic data type inference and customization during loading
Efficient loading and memory management for large datasets

2. Powerful Data Manipulation Features

Intuitive data manipulation with DataFrames and Series
SQL-style data joining (JOIN) operations
Missing value handling and imputation
Duplicate data detection and removal

3. Flexible Aggregation and Grouping

Multi-dimensional aggregation using GroupBy
Time series data aggregation and resampling
Pivot table creation
Rich statistical calculation functions

By effectively utilizing Pandas, you can implement complex data processing with concise code and build analysis functions needed for BI dashboards.

Plotly for Creating Charts

Plotly is one of the most important libraries in Streamlit BI dashboard development. It's suitable for data visualization because it easily creates interactive and beautiful graphs. Think of it as processing data with Pandas and visualizing with Plotly.

Plotly Features

1. Interactive Graphs

Operations like zoom, pan, and hover are possible
Users can intuitively explore data
Responsive design supporting various devices

2. Rich Graph Types

Basic line charts, bar charts, scatter plots
Specialized statistical graphs (heatmaps, box plots)
Map visualizations for geographic data
3D graphs and animations

3. Seamless Integration with Streamlit

Easy graph display with st.plotly_chart()
Responsive support with use_container_width=True
Can integrate with other Streamlit components

By effectively utilizing Plotly, you can build BI dashboards that intuitively convey data insights.

Review

This chapter taught you how to use Claude Code for basic dashboard construction.

The next chapter covers error handling. Error codes are a major barrier for programming beginners, but with proper understanding and Claude Code's help, they become excellent learning opportunities.

On this page