91
SQL reserved words – An empirical list
These words cause problems as table or column name
92
CTEs as lookup tables
How to use a CTE instead of complicated CASE expressions
93
Exploratory data analysis for humanities data
The course was very much an experiment, an attempt to combine literary study with computing. The 17 surviving students were largely sophomores. About two thirds of them were hard-core humanities majors; the rest were potential or actual CS majors or other STEM concentrators. The class met once a week for three hours…
94
95
Ban 1+N in Django
I always thought of 1+N as a thing that you just keep in your head, catch on code reviews or via performance regressions. This worked well for a long …
96
A baffling scale transform on a chart of university course selection trends
Harvard Magazine printed the following chart that confirms a trend in undergraduate majors that we all know about: students are favoring STEM majors at the expense of humanities.
97
A Relational Spreadsheet
I tried doing an INNER JOIN in Google Sheets and it was so awkward I was compelled to build my own lightweight relational spreadsheet / language / system.
98
99
100
How to implement Q&A against your docs with GPT3 embeddings and Datasette
If you’ve spent any time with GPT-3 or ChatGPT, you’ve likely thought about how useful it would be if you could point them at a specific, current collection of text …
101
SQLite 3.44: Interactive release notes
Order by in aggregates, better string concatenation, and thorough date formatting.
102
Using command line to process CSV files (2022)
Process CSV files from the command line with awk, sort, and grep: pull out fields, order rows, and filter data without a spreadsheet or a script.
103
Most Bitcoin Inscriptions belong to a single person
TLDR We have analyzed all Ordinal Inscriptions created since Bitcoin block 779832, which marks the introduction of the BRC-20 standard.
104
Trying chDB, an embeddable ClickHouse engine
A lightweight in-process SQL OLAP engine, powered by ClickHouse.
105
There have been 562 bank failures since 2000
One way to measure the magnitude of a bank's failure is by the amount of assets the bank held. At the time it failed, SVB's collapse was the second largest American bank failure of all time, and the largest since 2008.
106
107
Antirez' tiny JSON selector library
Salvatore Sanfilippo, aka antirez released a telegram bot library in C recently. While I'm not at all interested in a telegram bot library, I always like to read Salvatore's code, which is uniformly superb.
108
109
110
Mapping hospital accessibility with OpenStreetMap
Using OpenStreetMap data to craft a self-hosted interactive map of hospital accessibility in Virginia.
111
Statistical Arbitrage – An Easy Walkthrough
Statistical arbitrage (stat arb) is a pillar of quantitative trading that relies on mean reversion to predict the future returns of an asset. Mean reversion believes that if a stock has risen higher it’s more likely to revert in the short term which is the opposite of a momentum strategy that believes if a stock has…
112
UCLA professor refuses to cover for Dan Ariely in issue of data provenance
In one of the more bizarre twists of the Dan Ariely saga, there is a huge issue of provenance in one of Dan Ariely’s studies. The study in question is Dan Ariely’s Ten Commandments study at JMR. The hugely impactful paper has over 3,000 citations but unfortunately failed to replicate in a large-scale systematic…
113
Our humble attempt at “how much data do you need to fine-tune”
Efficacy of OpenAI Fine-tuning API, Cost/Latency Considerations, Experiencing Catastrophic Forgetting, and...Getting Roasted by a Fine-tuned GPT-3.5
114
We're wasting money by only supporting gzip for raw DNA files
Stylised graphics of money with DNA helix in the foreground.
115
The Architecture of Serverless Data Systems
I recently blogged about why I believe the future of cloud data services is large-scale and multi-tenant, citing, among others, S3. “ Top tier SaaS services like S3 are able to deliver amazing simplicity, reliability, durability, scalability, and low price because their technologies are s
116
Summing columns in remote Parquet files using DuckDB
vivym/midjourney-messages on Hugging Face is a large (~8GB) dataset consisting of 55,082,563 Midjourney images - each one with the prompt and a URL to the image hosted on Discord.
117
118
The new local data stack: Integrating Dagster, dbt and DuckDB
Lean and efficient MDS experience: Delivers better software engineering practices to the data ecosystem with the new local MDS stack comprised of Dagster, dbt and DuckDB which offers better developer productivity by enhancing testability of the E2E pipeline.
119
Short, friendly base32 slugs from timestamps
~4 lines of Go to produce succinct, stable, URL-friendly slugs from timestamps.
120
Demystifying Text Data with the Unstructured Python Library
In the world of data, textual data stands out as being particularly complex. It doesn’t fall into neat rows and columns like numerical data does. As a side project, I’m in the process of developing my own personal AI assistant. The objective is to use the data within my notes and documents to answer my questions. The…