Car Sales Data Engineering Analytics

reason-machines/data-skills/skills/car-sales-data-engineering-analytics

作者 reason-machines8e7bd5f22574無授權條款4 個星標收錄於 2026年10月8日更新於 2026年10月8日儲存庫2 個月前更新

Process, clean, and analyze car sales data with statistical modeling and interactive Streamlit dashboards for business insights.

僅含說明Data & Analytics
AI 產生的概覽

處理、清理並統計分析汽車銷售資料，產生圖表與互動式 Streamlit 儀表板。

功能
此技能描述一套針對約 2.4 萬筆汽車銷售紀錄的 ETL 與分析框架。內容涵蓋載入與清理 CSV、計算描述性統計與營收指標、執行 15 項預設分析（例如價格分布、地區與月度趨勢、t 檢定、多元線性迴歸、Z 分數異常值偵測與 Shapiro-Wilk 常態性檢定），並儲存清理後的資料與 PNG 圖表。它也說明具備篩選、分群比較與 CSV 匯出功能的 Streamlit 儀表板。
適用情境
適用於處理汽車銷售資料集，且需要清理、統計建模或商業導向視覺化時。適合價格、地區、人口統計或產品組合分析，以及建置互動式銷售儀表板的需求。
執行需求
需要 Python 3.10+ 以及 pandas、numpy、matplotlib、scipy 和 streamlit，並以 uv 管理。需要汽車銷售 CSV 檔案，並會輸出清理後的資料與圖表。此技能僅包含說明，未附帶指令碼。

Car Sales Data Engineering & Analytics

Skill by ara.so — Data Skills collection.

A comprehensive data engineering and analytics framework for processing ~24K car sales records with ETL pipelines, statistical modeling, and interactive Streamlit dashboards. Provides 15 pre-built analyses covering pricing trends, regional patterns, demographic insights, and feature correlations.

Installation

This project uses uv for package management:

bash
# Clone the repositorygit clone https://github.com/Abdumalik-ProDev/Car-Sales-Data-Engineering.gitcd Car-Sales-Data-Engineering
# Install dependenciesuv sync

Dependencies: Python 3.10+, pandas, numpy, matplotlib, scipy, streamlit

Quick Start

Launch Interactive Dashboard

bash
# Start Streamlit web UIuv run streamlit run src/ui.py
# Alternative via entry pointuv run python -m src.main

Run Full Pipeline

bash
# Execute all 15 analyses and generate figuresuv run python -m src.main --pipeline

This will:

  • Load and clean data/Car sales.csv
  • Generate outputs/cleaned_data.csv
  • Create 15 PNG charts in outputs/figures/

Core Module: src/analysis.py

The main analysis engine provides ETL, statistics, and modeling capabilities.

Data Loading & Cleaning

python
from src.analysis import CarSalesAnalysis
# Initialize analyzeranalyzer = CarSalesAnalysis('data/Car sales.csv')
# Access cleaned datadf = analyzer.dataprint(f"Total records: {len(df)}")print(f"Columns: {df.columns.tolist()}")
# Save cleaned datasetanalyzer.save_cleaned_data('outputs/cleaned_data.csv')

Key Columns:

  • car_id, date, customer_name, dealer_name, company, model
  • year, price, body_style, transmission, color
  • dealer_no, dealer_region, phone, gender, annual_income

Statistical Summaries

python
# Get descriptive statisticsstats = analyzer.describe_data()print(stats)
# Revenue metricstotal_revenue = analyzer.data['price'].sum()avg_price = analyzer.data['price'].mean()median_price = analyzer.data['price'].median()
print(f"Total Revenue: ${total_revenue:,.0f}")print(f"Avg Price: ${avg_price:,.0f}")print(f"Median Price: ${median_price:,.0f}")

Generate Individual Analyses

python
# Q1: Price distributionanalyzer.plot_price_distribution(save_path='outputs/figures/q1_price_dist.png')
# Q2: Monthly sales trendanalyzer.plot_monthly_sales_trend(save_path='outputs/figures/q2_monthly_trend.png')
# Q3: Sales by regionanalyzer.plot_sales_by_region(save_path='outputs/figures/q3_regional_sales.png')
# Q6: Income vs Price regressionanalyzer.plot_income_vs_price(save_path='outputs/figures/q6_income_price.png')
# Q9: Automatic vs Manual transmission comparison (t-test)analyzer.compare_transmission_prices(save_path='outputs/figures/q9_transmission.png')

Statistical Modeling

python
# Q12: Multiple linear regression# Predicts price from year, annual_income, transmissionanalyzer.multiple_regression_analysis(save_path='outputs/figures/q12_regression.png')
# Q13: Detect outliers using Z-scoresanalyzer.detect_outliers_zscore(save_path='outputs/figures/q13_outliers.png')
# Q15: Test price normality with Shapiro-Wilkanalyzer.test_normality(save_path='outputs/figures/q15_normality.png')

Streamlit Dashboard (src/ui.py)

Page Structure

The dashboard provides 6 interactive sections:

  1. 📊 Overview - Data summary, sample rows, statistics
  2. 💰 Sales & Revenue - Price trends, regional analysis
  3. 👥 Demographics - Gender, income patterns
  4. 🔧 Product Insights - Brand, body style, transmission
  5. 📈 Statistical Modeling - Regression, outliers, normality
  6. 🔍 Filter & Explore - Custom filters with CSV export
  7. ⚖️ Compare Segments - Side-by-side comparison with t-tests

Custom Filtering Example

python
# Users can filter via sidebar widgets# Example: Filter cars by price range and region
# In ui.py, the filter logic:filtered = analyzer.data.copy()
if price_range:    filtered = filtered[        (filtered['price'] >= price_range[0]) &         (filtered['price'] <= price_range[1])    ]
if selected_regions:    filtered = filtered[filtered['dealer_region'].isin(selected_regions)]
if selected_companies:    filtered = filtered[filtered['company'].isin(selected_companies)]
# Display and exportst.dataframe(filtered)st.download_button(    "Download CSV",    filtered.to_csv(index=False),    "filtered_sales.csv")

Common Analysis Patterns

Price Analysis by Category

python
# Average price by car companycompany_prices = analyzer.data.groupby('company')['price'].mean().sort_values(ascending=False)print(company_prices.head(10))
# Price by body stylebody_prices = analyzer.data.groupby('body_style')['price'].agg(['mean', 'median', 'count'])print(body_prices)
# Price by transmission typetrans_prices = analyzer.data.groupby('transmission')['price'].describe()print(trans_prices)

Regional & Temporal Analysis

python
# Sales volume by regionregional_sales = analyzer.data['dealer_region'].value_counts()print(regional_sales)
# Monthly revenue trendanalyzer.data['month'] = pd.to_datetime(analyzer.data['date']).dt.to_period('M')monthly_revenue = analyzer.data.groupby('month')['price'].sum()print(monthly_revenue)
# Year-over-year comparisonyearly_sales = analyzer.data.groupby('year').agg({    'price': ['sum', 'mean', 'count']})print(yearly_sales)

Statistical Tests

python
from scipy import stats
# Compare prices: Automatic vs Manual transmissionauto_prices = analyzer.data[analyzer.data['transmission'] == 'Automatic']['price']manual_prices = analyzer.data[analyzer.data['transmission'] == 'Manual']['price']
t_stat, p_value = stats.ttest_ind(auto_prices, manual_prices)print(f"T-statistic: {t_stat:.4f}, P-value: {p_value:.4f}")
# Correlation between income and pricecorrelation = analyzer.data['annual_income'].corr(analyzer.data['price'])print(f"Income-Price Correlation: {correlation:.4f}")

Configuration

File Paths

Default paths are defined in src/analysis.py:

python
# Customize data pathsanalyzer = CarSalesAnalysis('custom_path/sales_data.csv')
# Custom output directoryanalyzer.save_cleaned_data('custom_output/cleaned.csv')
# Figures directoryos.makedirs('custom_figures', exist_ok=True)analyzer.plot_price_distribution(save_path='custom_figures/prices.png')

Streamlit Configuration

Create .streamlit/config.toml for dashboard customization:

toml
[theme]primaryColor = "#FF4B4B"backgroundColor = "#FFFFFF"secondaryBackgroundColor = "#F0F2F6"textColor = "#262730"
[server]port = 8501headless = trueenableCORS = false

Running Full Pipeline Programmatically

python
from src.analysis import CarSalesAnalysisimport os
# Initializeanalyzer = CarSalesAnalysis('data/Car sales.csv')
# Create output directoriesos.makedirs('outputs/figures', exist_ok=True)
# Save cleaned dataanalyzer.save_cleaned_data('outputs/cleaned_data.csv')
# Generate all 15 analysesanalyses = [    ('q1_price_dist.png', analyzer.plot_price_distribution),    ('q2_monthly_trend.png', analyzer.plot_monthly_sales_trend),    ('q3_regional_sales.png', analyzer.plot_sales_by_region),    ('q4_gender_split.png', analyzer.plot_gender_distribution),    ('q5_income_region.png', analyzer.plot_income_by_region),    ('q6_income_price.png', analyzer.plot_income_vs_price),    ('q7_company_prices.png', analyzer.plot_avg_price_by_company),    ('q8_body_style.png', analyzer.plot_price_by_body_style),    ('q9_transmission.png', analyzer.compare_transmission_prices),    ('q10_colors.png', analyzer.plot_popular_colors),    ('q11_heatmap.png', analyzer.plot_body_transmission_heatmap),    ('q12_regression.png', analyzer.multiple_regression_analysis),    ('q13_outliers.png', analyzer.detect_outliers_zscore),    ('q14_dealer_prices.png', analyzer.plot_dealer_prices),    ('q15_normality.png', analyzer.test_normality),]
for filename, func in analyses:    func(save_path=f'outputs/figures/{filename}')    print(f"✓ Generated {filename}")

Troubleshooting

Missing Data Issues

python
# Check for missing valuesmissing = analyzer.data.isnull().sum()print(missing[missing > 0])
# Handle missing valuesanalyzer.data = analyzer.data.dropna(subset=['price', 'year'])analyzer.data['annual_income'].fillna(analyzer.data['annual_income'].median(), inplace=True)

Date Parsing Errors

python
# Ensure proper date formatanalyzer.data['date'] = pd.to_datetime(analyzer.data['date'], errors='coerce')analyzer.data = analyzer.data.dropna(subset=['date'])

Memory Issues with Large Datasets

python
# Load only required columnsusecols = ['price', 'company', 'body_style', 'dealer_region', 'year']df = pd.read_csv('data/Car sales.csv', usecols=usecols)
# Use dtype optimizationdf['price'] = df['price'].astype('float32')df['year'] = df['year'].astype('int16')

Streamlit Port Conflicts

bash
# Specify custom portuv run streamlit run src/ui.py --server.port 8502
# Or in configecho "[server]\nport = 8502" > .streamlit/config.toml

Key Insights Reference

  • Total Records: 23,906 sales
  • Revenue: $655.6M total
  • Pricing: $27,426 avg, $23,000 median
  • Top Body Style: SUV (27%)
  • Top Region: Austin (17%)
  • Premium Brand: Cadillac ($37,557 avg)
  • Demographics: 79% Male, 21% Female
  • Transmission: 53% Automatic, 47% Manual

來源與署名

來源：reason-machines/data-skills位於skills/car-sales-data-engineering-analytics提交8e7bd5f

授權條款： 無授權條款

內容歸原作者所有。SourceWeft 從公開儲存庫中收錄這些內容。

檢舉或申請下架

更多來自 reason-machines/data-skills 的技能

Youtube Channel Analytics Dashboard

reason-machines

介紹一款用於追蹤 YouTube 頻道觀看數、訂閱者、收益與受眾資料的 Windows 儀表板工具。

Data & Analytics42 個月前更新

Terraform Iac Data Engineering

reason-machines

提供用於在 AWS 上佈建 S3、EC2 和 IAM 等資料工程資源的 Terraform 基礎設施即程式碼模式。

DevOps & Cloud42 個月前更新

Terraform Data Engineering Infrastructure

reason-machines

使用 Terraform 為資料工程佈建 AWS 基礎設施（S3、EC2、IAM）的模式與指令。

DevOps & Cloud42 個月前更新

Retail Etl Pipeline Medallion

reason-machines

指導使用 SQL Server 與選用 PySpark 建立包含青銅、白銀、黃金三層的零售 ETL 資料倉儲。

Data & Analytics42 個月前更新

Power Bi Salespulse 360 Dashboard

reason-machines

指導建置 SalesPulse 360 Power BI 零售儀表板，用於銷售、利潤、預測與警示分析。

Data & Analytics42 個月前更新

Logifleet Pulse Supply Chain Analytics

reason-machines

在 MS SQL Server 上建置物流資料倉儲，並以 Power BI 儀表板分析倉儲與車隊資料。

Data & Analytics42 個月前更新

更多Data & Analytics技能

Swap Curve Strategy

anthropics

精選

運用 MCP 定價工具分析利率交換曲線、交換利差、實質利率與曲線交易機會。

Data & Analytics39K2 週前更新

Unit Economics

anthropics

精選

為私募股權標的分析單位經濟：ARR 世代、LTV/CAC、淨留存、回收期與利潤瀑布。

Data & Analytics39K2 週前更新

Option Vol Analysis

anthropics

精選

結合波動率曲面、含希臘字母的選擇權定價與歷史價格，比較隱含波動率與已實現波動率。

Data & Analytics39K2 週前更新

Macro Rates Monitor

anthropics

精選

運用 MCP 的 GDP、通膨、殖利率曲線、損益兩平通膨與交換利率資料，建立總體與利率儀表板。

Data & Analytics39K2 週前更新

Clean Data Xls

anthropics

精選

清理雜亂的試算表資料：修剪空白、統一大寫、轉換文字數字、標準化日期並移除重複項。

Data & Analytics39K2 週前更新

Fixed Income Portfolio

anthropics

精選

透過債券定價、彙總風險指標、預測現金流量並執行利率情境，審查固定收益投資組合。

Data & Analytics39K2 週前更新