cartodb4/crankshaft
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Merge pull request #155 from CartoDB/update-markov-docs-null

Browse Source 
adds caveats about usage
This commit is contained in:
Andy Eschbacher
2018-01-09 11:22:10 -05:00
committed by GitHub
parent f1428a3e36 8bbfac0dbc
commit 4ece712cae
2 changed files with 18 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
doc/02_moran.md
View File

@@ -1,5 +1,14 @@
## Areas of Interest Functions
A family of analyses to uncover groupings of areas with consistently high or low values (clusters) and smaller areas with values unlike those around them (outliers). A cluster is labeled by an 'HH' (high value compared to the entire dataset in an area with other high values), or its opposite 'LL'. An outlier is labeled by an 'LH' (low value surrounded by high values) or an 'HL' (the opposite). Each cluster and outlier classification has an associated p-value, a measure of how significant the pattern of highs and lows is compared to a random distribution.
These functions have two forms: local and global. The local versions classify every input geometry while the global function gives a rating of the overall clustering characteristics of the dataset. Both forms accept an optional denomiator (see the rate versions) if, for example, working with count data and a denominator is needed.
### Notes
* Rows with null values will be omitted from this analysis. To ensure they are added to the analysis, fill the null-valued cells with an appropriate value such as the mean of a column, the mean of the most recent two time steps, or use a `LEFT JOIN` to get null outputs from the analysis.
* Input query can only accept tables (datasets) in the users database account. Common table expressions (CTEs) do not work as an input unless specified within the `subquery` argument.
### CDB_AreasOfInterestLocal(subquery text, column_name text)
This function classifies your data as being part of a cluster, as an outlier, or not part of a pattern based the significance of a classification. The classification happens through an autocorrelation statistic called Local Moran's I.
@@ -29,6 +38,7 @@ A table with the following columns.
| vals | NUMERIC | Values from `'column_name'`. |
#### Example Usage
```sql

9
doc/04_markov.md
View File

@@ -8,7 +8,7 @@ This function takes time series data associated with geometries and outputs like
| Name | Type | Description |
|------|------|-------------|
| subquery | TEXT | SQL query that exposes the data to be analyzed (e.g., `SELECT * FROM real_estate_history`). This query must have the geometry column name `the_geom` and id column name `cartodb_id` unless otherwise specified in the input arguments |
| subquery | TEXT | SQL query that exposes the data to be analyzed (e.g., `SELECT * FROM real_estate_history`). This query must have the geometry column name `the_geom` and id column name `cartodb_id` unless otherwise specified in the input arguments. Tables in queries must exist in user's database (i.e., no CTEs at present) |
| column_names | TEXT Array | Names of column that form the history of measurements for the geometries (e.g., `Array['y2011', 'y2012', 'y2013', 'y2014', 'y2015', 'y2016']`). |
| num_classes (optional) | INT | Number of quantile classes to separate data into. |
| weight type (optional) | TEXT | Type of weight to use when finding neighbors. Currently available options are 'knn' (default) and 'queen'. Read more about weight types in [PySAL's weights documentation](https://pysal.readthedocs.io/en/v1.11.0/users/tutorials/weights.html). |
@@ -30,12 +30,19 @@ A table with the following columns.
| rowid | NUMERIC | id of the row that corresponds to the `id_col` (by default `cartodb_id` of the input rows) |
#### Notes
* Rows will null values will be omitted from this analysis. To ensure they are added to the analysis, fill the null-valued cells with an appropriate value such as the mean of a column, the mean of the most recent two time steps, etc.
* Input query can only accept tables (datasets) in the users database account. Common table expressions (CTEs) do not work as an input unless specified in the `subquery` parameter.
#### Example Usage
```sql
SELECT
c.cartodb_id,
c.the_geom,
c.the_geom_webmercator,
m.trend,
m.trend_up,
m.trend_down,