Project 05 · Portfolio Dashboard · local

My money was spread across five inboxes. Could one script bring it home?

Fund statements arrive as password-locked PDFs, the broker lives behind an API, and prices live somewhere else again. This pipeline stitches it together once a month, entirely on my laptop.

Focus: The architecture

Source

Step 1 of 6

It starts in the inbox.

Fund managers email monthly & quarterly statements as PDFs. A script searches Gmail (read-only access) for exactly those senders & subjects.

Step 2 of 6

Unlock & read the PDFs.

Each statement is password-locked. The script unlocks it, then pdfplumber pulls the numbers out of the tables: value, invested, withdrawn, holdings.

Step 3 of 6

Everything lands in one database.

A local SQLite file. One table per source, one row per month, so history builds up on its own.

Step 4 of 6

Add the broker & the market.

Direct holdings come from the broker's API, and daily price history from a market-data library. Same database.

Step 5 of 6

One dashboard on top.

Streamlit & Plotly turn it into charts: total value over time, returns by source, allocation. One command refreshes the lot each month.

Step 6 of 6

And none of it leaves the laptop.

No cloud database, no hosted app, no third party holding my numbers. That's a design choice, which is why this page runs on made-up data.

Sources: Gmail, PDFs, broker API, prices
Store: SQLite
Refresh: Monthly, one command
Cloud: None, on purpose

Built with Python · Gmail API · pdfplumber · SQLite · Streamlit

Next project

LBS Yearbook →

Post moments as they happen, sign each other's pages, and get a 60-second Wrapped of your year.

Setting the type…

The Yash Hour