Developers drift into finance the same way they drift into anything else: someone mentions a trading bot, a market dashboard or a DeFi trick, and an evening disappears into docs. Over the past month I saved notes on eight resources that kept coming up in that drift, from open-source trading and data tools to Solana and Ethereum primers. Rereading them side by side, my takeaway is that the code is rarely the hard part. The hard part is knowing what risk you're actually taking, and the better projects below say so loudly. None of this is investment advice; treat it as a reading list for builders.
Tools for trading and market data
Freqtrade. This is the open-source crypto trading bot most people start with, and for good reason. It's written in Python, keeps state in SQLite, and gives you backtesting, plotting and hyperparameter search (hyperopt) before a single real order goes out, with FreqAI for adaptive ML-based strategies on top. You can drive it from Telegram or a built-in web UI, and it talks to most large exchanges for spot and futures. What I like most is the honesty of its docs: run in dry-run mode first, read the source, and assume you're responsible for every loss. My notes on Freqtrade's backtest-first workflow cover the CLI commands and the modest hardware it needs.
OpenBB. OpenBB has shifted from "open-source terminal" to something more like a data plumbing layer. Its Open Data Platform lets you wire proprietary, licensed and public data sources in once, then consume them from Python (pip install openbb), a local FastAPI server, Excel, an analyst-facing Workspace UI, or an MCP server for AI agents. That last surface is the interesting one: the same normalized price history a quant pulls in a notebook can be handed to an agent without a second integration. Here's how OpenBB's Open Data Platform is structured, including its Apache-2.0 license and its warning that the data isn't guaranteed to be accurate.
A market heatmap for China A-shares. Sometimes you don't want a bot, you want a picture. 大盘云图 is a free web heatmap that lays out every Shanghai and Shenzhen listed stock by industry sector, colours each tile by its daily move, and lets you zoom from sector to single ticker, with Hong Kong indices on top. It refreshes roughly every eight seconds during trading hours from mainstream delayed feeds (not Level-2), needs no login, and supports full-screen monitoring and replaying past sessions. If you've used Finviz-style maps for US stocks, this A-share heatmap tool is the equivalent view of the Chinese market.
A four-quadrant allocation framework. This one is a personal system rather than software: a self-described engineer splits money along two axes, volatility and time horizon, and gives each quadrant its own tactics. The bulk sits in low-volatility, short-term positions such as stablecoin yield, with smaller slices for long-term index-fund dollar-cost averaging, trend-following quant strategies on assets like BTC and big-tech stocks, and a deliberately tiny high-risk bucket the author openly says has a low win rate. The page reports strong backtest numbers for an ETF rotation strategy, which deserve the usual scepticism about backtests. The framing is still worth borrowing, and the note sums up the volatility-by-horizon investment quadrants.
Learning the on-chain side
Flash loans, explained. Flash loans are the DeFi feature that sounds impossible until you understand transaction atomicity. You borrow with no collateral, use the funds, and repay inside the same Ethereum transaction; if repayment fails, the whole transaction reverts as if the loan never happened. That makes them useful for arbitrage and, unfortunately, for amplifying protocol exploits. The note also flags the less dramatic costs: you usually have to write and deploy your own Solidity contract, and you pay gas even when the trade fails. A clear primer: what flash loans are and why they're tied to so many DeFi hacks.
A Python programmer's route into blockchain. If Solidity feels like a leap, this guide starts where Python developers already are. It walks from distributed ledgers and proof-of-work versus proof-of-stake, to building a toy blockchain with hashing and chain validation, adding transactions, mining and a Flask API, and finally talking to Ethereum through Web3.py and Remix. It ends with a reading list and a candid section on energy use, governance and regulation. Good for a weekend: a blockchain learning path for Python developers.
pxsol: learning Solana from the ground up. pxsol's Chinese-language book is one of the more thorough Solana walkthroughs I've come across. It goes from keys, Base58 addresses and vanity-address scams, through hand-built transactions, the account model, program-derived addresses and rent exemption, to writing and upgrading Rust programs, issuing SPL tokens and designing airdrops, and closes with Solana's economics and tooling (Anchor, Pinocchio, web3.js, solana-py). The note maps the chapters of the pxsol Solana guide so you can jump to the part you need.
From token to trading pool. One chapter of that series deserves its own entry because it's where theory meets a live market. It shows how the example token goes from mainnet issuance to a Raydium pool paired with SOL: choosing the fee tier and initial price, receiving LP tokens that represent your share of the pool, and then watching what happens when that pool is small. The lesson about slippage in thin liquidity is one every token launcher should learn on paper first. See the Solana DEX listing walkthrough.
Put together, these eight form a rough curriculum: understand the mechanics (flash loans, Python-built chains, Solana from first principles), get decent data (OpenBB, a heatmap for a quick read), test ideas without real money (Freqtrade's dry runs), and decide how much risk you're willing to hold before you write a line of strategy code. The tools have never been easier to run. Knowing when not to run them is still on you. If you've taken any of these from experiment to real capital, I'd like to hear what surprised you.
Top comments (0)