Ledgr

Lightweight serverless expense tracker. Requires: iOS Shortcuts + AWS + Grafana Cloud (Free Tier)

Why I Built This

I needed a quick and easy way to track my monthly spending and answer questions like:

• "What have I spent in total this month?"
• "What am I spending my money on?"
• "How many times per month am I going out to purchase things?"
• "What am I spending the most money on?"
• "How many times a month am I spending money at bars or cafes?"
• "What are my spending trends?"
• "Where can I cut back on spending?"

Goals

1. Log expenses from iPhone lock screen in under 5 seconds
2. Build a dashboard to view monthly spending, category breakdowns, insights into top vendors, and trends.

Architecture

Logging Flow iOS Shortcut --> API Gateway --> DynamoDB --> Streams --> Lambda(Categorize & Sum) --> DynamoDB

Visualization Flow Grafana --> Athena --> Lambda (DB Connector) --> DynamoDB

DynamoDB Structure

id (n): YYYYMM (Primary Key)
timestamp (n): $context.requestTimeEpoch (Sort Key)
---
category (s): Category of expense. When pulled via Lambda, please python.string.upper()
description (s): Likely the name of the vendor
amount (n): Value of expense

iOS Shortcut Setup

1. Install AWS SAM CLI
2. Download template.yaml
3. Run sam deploy --guided and follow the prompts (replace variable defaults)
4. Get API Key from API Gateway Console
5. Create iOS Shortcut
   • List: All expense categories that are relevant
   • Choose from List for Category
   • Ask for 'Text' for Description
   • Ask for 'Number' for Amount
   • Get Current Date and format it to yyyyMMddHHmmss. I need to change this as HHmmss is unnecessary.
   • Set URL to the SAM output 'ApiBaseUrl' + /log --> Something like https://<api_id>.execute-api.<aws_region>.amazonaws.com/dev/log
   • Get Contents of URL
     • Headers: x-api-key: <your_api_key>
     • Body: DAY: <formatted_date>, CATEGORY: <select_from_variable>, DESCRIPTION: <ask_for_input>, AMOUNT: <ask_for_input>
   • Get 'Value' for message in Contents of URL
   • Show Output
6. Put iOS Shortcut on lockscreen

Athena Setup: Athena Federated Query via DynamoDB Connector

1. Deploy DynamoDB Connector

aws serverlessrepo create-cloud-formation-change-set \
  --application-id arn:aws:serverlessrepo:us-east-1:292517598671:applications/AthenaDynamoDBConnector \
  --stack-name athenaDynamoDBConnector \
  --capabilities CAPABILITY_RESOURCE_POLICY CAPABILITY_IAM \
  --parameter-overrides '[{"Name":"AthenaCatalogName","Value":"dynamo"},{"Name":"SpillBucket","Value":"karl-macro-tracker-athena-spill-bucket"}]'

2. Register Connector as Data Catalog in Athena

ACCOUNT_ID=$(aws sts get-caller-identity --query Account --output text)
REGION=$(aws configure get region)

aws athena create-data-catalog \
  --name dynamo \
  --type LAMBDA \
  --description "DynamoDB Federated Connector" \
  --parameters function=arn:aws:lambda:${REGION}:${ACCOUNT_ID}:function:dynamo

3. Create Dashboard & Panels in Grafana

For large tables queried frequently, it's often cheaper to export DynamoDB data to S3 (via DynamoDB Streams or point-in-time exports) and query from there with Athena at standard S3 rates --> Evaluate later to determine if this is necessary. Check Cost Explorer

---

Todo

1. DONE! Month-over-month change: % change from last month
2. NEED MORE DATA! Rolling 3 month average
3. Line chart for category trends over time
4. Top vendors per category
5. Update queries to use Grafana's Time Range feature (macros)