Introduction to Spells
Historical administrative data for family self-sufficiency programs is usually tracked as a series of transactions or point in time data. An alternative method of representing the same data is spells. Each spell defines a start and end date for a particular set of characteristics. It is frequently convient to store data in a spells format for analysis. The code in the section will introduce you to converting point in time data to spells data.
Point in Time Data
Each row in point in time data includes one date and several other characteristics such as case id and benefit status.
| case_id | date | status |
|---|---|---|
| 578 | 2001-01-01 | 1 |
| 578 | 2001-02-01 | 0 |
| 578 | 2001-03-01 | 1 |
| 933 | 2001-01-01 | 1 |
| 933 | 2001-02-01 | 1 |
| 933 | 2001-03-01 | 1 |
Spells
Each row in a spells file includes a start date and an end date and several other characteristics such as case id and benefit status.
| case_id | start_date | end_date | status |
|---|---|---|---|
| 578 | 2001-01-01 | 2001-01-31 | 1 |
| 578 | 2001-02-01 | 2001-02-28 | 0 |
| 578 | 2001-03-01 | 2001-03-31 | 1 |
| 933 | 2001-01-01 | 2001-03-31 | 1 |
General Spell Creation
We have provided sample code for creating spells in R and Python.