|
| 1 | +Information Management for Journalists |
| 2 | +====================================== |
| 3 | + |
| 4 | +Introducing data |
| 5 | +---------------- |
| 6 | + |
| 7 | +* Types of data: |
| 8 | + * Simple types |
| 9 | + * Numbers (bytes, ints, floats) |
| 10 | + * Text (chars, strings) |
| 11 | + * Boolean |
| 12 | + * Enumerations |
| 13 | + * Binary blobs (images) |
| 14 | + * Complex types |
| 15 | + * Lists (1D) |
| 16 | + * Tables (2D) |
| 17 | + * Relationships (3 or more dimensions) |
| 18 | + * Objects (hierarchical data) |
| 19 | +* Metadata: |
| 20 | + * Unique ids |
| 21 | + * Constraints ("business rules") |
| 22 | + |
| 23 | + |
| 24 | +Storing data |
| 25 | +------------ |
| 26 | + |
| 27 | +* Lists: Plain text |
| 28 | +* Tables: Spreadsheets (CSV, Excel, Google Docs) |
| 29 | +* Relationships: SQL Databases (MySQL, PostgreSQL, sqlite) |
| 30 | +* Objects: Flat-files (JSON, YAML) or Object Databases (MongoDB, CouchDB) |
| 31 | + |
| 32 | + |
| 33 | +Tracking changes |
| 34 | +---------------- |
| 35 | + |
| 36 | +* What constitutes a change in version? |
| 37 | +* The timestamp: atomic unit of change |
| 38 | +* Sometimes it makes sense to record that things don't change too |
| 39 | +* Files and folders (a.k.a naming things) |
| 40 | + * Always name from less specific to more specific |
| 41 | + * Always timestamp things you're changing |
| 42 | + * Always make copies (you have unlimited hard drive space--use it) |
| 43 | +* Dropbox: The poor man's version control |
| 44 | +* git: The engineer's version control |
| 45 | + * git in 5 minutes |
| 46 | + * Commit messages matter |
| 47 | + * Works for text formats (CSV, flat files), but not for binaries |
| 48 | + * Not designed for big data |
| 49 | +* Change tables and data warehouses: creating problems to solve problems |
| 50 | + |
| 51 | + |
| 52 | +Data pipelines |
| 53 | +-------------- |
| 54 | + |
| 55 | +* Why repeatable processes? |
| 56 | + * Ensure clear provenance of the data |
| 57 | + * Depend against attacks on integrity |
| 58 | + * Self-documenting workflow |
| 59 | + * Show your work |
| 60 | +* Acquisition |
| 61 | + * Ways to get data |
| 62 | + * Curl |
| 63 | + * APIs |
| 64 | + * Scrapers |
| 65 | + * Documenting provenance |
| 66 | +* Processing |
| 67 | + * Scripts |
| 68 | + * Bash scripts |
| 69 | + * csvkit |
| 70 | + * Python |
| 71 | + * Documenting order of operations |
| 72 | + * Naming things; revisited |
| 73 | +* Shims, adapters and glue code |
| 74 | +* Crons |
| 75 | + |
0 commit comments