GitHub Traffic Analytics

Track GitHub repository traffic (views and clones) over time. GitHub only exposes the last 14 days; this tool captures data each day, appends it to a central CSV in your storage account, and preserves the history so you can analyze older traffic for free.

What It Does

• Fetches views and clones data for all your public repositories
• Stores historical data (GitHub only keeps 14 days)
• First run: Backfills 14 days of historical data (T-15 through T-2)
• Subsequent runs: Captures data from 2 days ago (T-2) daily, ensuring that the metrics are updated in GitHub (often there is a delay)
• Outputs: CSV with date columns showing views(clones) per repository. Sample output here.
• Visualize metrics in a PowerBI report, with filters to drill into a single repo or any selection

Usage Options

Get the code

# Clone and enter the repo
git clone https://github.com/claestom/github-traffic-analytics.git
cd github-traffic-analytics

Option 1: Azure Functions (automated)

Deploy to Azure for fully automated daily collection at 7 AM CET. On first run - when no CSV gets detected, the function backfills 14 days of history. Subsequent runs collect data from 2 days prior (T-2) to ensure data is captured before GitHub purges it.

Prerequisites:

• Azure subscription
• Azure CLI installed
• PowerShell 7.4+

Deploy:

# 1. Login to Azure
az login

Tip: copy-paste the script into a text editor, fill in the values, then run it as a whole from inside the github-traffic-analytics folder.

# 2. Set variables
$rg = "<rg-name>"
$location = "<region of deployment>"

# 3. Create resource group
az group create --name $rg --location $location

# 4. Deploy infrastructure (storage, function app, identity)
az deployment group create `
  --resource-group $rg `
  --template-file infra/main.bicep `
  --parameters infra/main.bicepparam

# 5. Get function app name from outputs
$funcApp = az functionapp list -g $rg --query "[0].name" -o tsv

# 6. Set GitHub credentials in the Azure Functions config
az functionapp config appsettings set -g $rg -n $funcApp `
  --settings GITHUB_TOKEN="ghp_new_token" GITHUB_USERNAME="your_username"

# 7. Publish function code
cd azure-function
$env:FUNCTIONS_WORKER_RUNTIME = "powershell"
$env:FUNCTIONS_WORKER_RUNTIME_VERSION = "7.4"
func azure functionapp publish $funcApp --nozip --powershell 

What Gets Created:

• Storage Account (stores CSV in metrics container)
• Function App (PowerShell 7.4, Consumption plan)
• User-Assigned Managed Identity (secure storage access)
• Application Insights (monitoring)

(optional) Once deployed:

Test the Azure Function:

Explained here as well.

Architecture:

Option 2: Local Script

Run manually or via scheduled task on your local machine.

Setup:

# 1. Create .env file in root directory
GITHUB_TOKEN=ghp_your_token_here
GITHUB_USERNAME=your_username

# 2. Run the script
cd src
.\github-traffic-metrics.ps1

Output: outputs/github-traffic-metrics.csv

Linux users:

sudo apt-get install -y powershell
pwsh ./src/github-traffic-metrics.ps1

Schedule (optional):

• Windows: Task Scheduler
• Linux/Mac: cron job

Power BI Usage (SAS + .pbit)

Use the template in powerbi/ to visualize the CSV with your own storage.

Generate a blob SAS (how to do it)

• Scope: blob-level, permissions: r (read only), set expiry date.

• Desktop setup:

  • Open the .pbit → enter the SAS URL when prompted.
  • Data source credentials: choose Anonymous (SAS is in the URL).
  • Refresh to apply all existing transform steps.

• Publish to Service:

  • Dataset → Settings → Data source credentials → set to Anonymous for the blob domain.
  • Enable Scheduled refresh (ensure SAS expiry is sufficient).

• SAS scope: prefer blob-level SAS with r permission only; rotate with brief overlap.

Common Use Cases

• Overall insights across repositories: use the line chart with Date on the X-axis and Views/Clones as values; leave the Repository slicer empty to see totals.
• Single repository focus: select one repo in the Repository slicer; visuals filter to that repo’s traffic only.
• Specific time period: use the Date slicer (Between) to limit visuals to a range (e.g., last 7 days).
• Example: Understand activities after a release window: pick the repo in the slicer and set the Date slicer to the release week to analyze impact.

Examples

Overall insights across repositories

Traffic for two repos for a specific time period

Analyze traffic after new release (use of time range and repo filter)

Cost Estimate

• Azure Functions (Consumption): ~30s daily run (up to 2 minutes on first run with backfill) stays in the free grant → ~$0.
• Storage (Blob, LRS, Hot): Few MB CSV + tiny egress → <$0.10/month.
• Power BI: Desktop free; sharing in Service needs Pro/PPU (only if you publish).

At low volume the stack is effectively free; costs grow with higher frequency, larger blobs, or publishing to Power BI Service.

More information about pricing: Azure services and Power BI.

Roadmap

Planned enhancements to deepen traffic insights:

• Unique Views & Clones: Track distinct visitors and unique IPs alongside total metrics to measure repeat engagement.
• Referring Websites: Capture traffic sources (Twitter, Reddit, blogs, search engines) to measure campaign impact.

The CSV-based approach ensures backward compatibility while unlocking richer insights.

Contributions

Contributions are always appreciated! Open an issue to report bugs or suggest features, or submit a PR to improve the solution.