Unlike most of the other data sources, the state-level ILINet data isn't updated automatically. The data was obtained around 2015-09-10 from the various states' websites, linked here: http://www.cdc.gov/flu/weekly/ This data was re-obtained around June 2016, and uploaded 2016-11-14. The process is lossy, and so the two version don't entirely agree. Because of this, we now also store a version tag with the data. The state_ili table stores this data:

+---------+-------------+------+-----+---------+----------------+
| Field   | Type        | Null | Key | Default | Extra          |
+---------+-------------+------+-----+---------+----------------+
| id      | int(11)     | NO   | PRI | NULL    | auto_increment |
| epiweek | int(11)     | NO   | MUL | NULL    |                |
| state   | varchar(12) | NO   | MUL | NULL    |                |
| ili     | double      | NO   |     | NULL    |                |
| version | int(11)     | NO   | MUL | NULL    |                |
+---------+-------------+------+-----+---------+----------------+

• id: unique identifier for each record
• epiweek: the epiweek during which the data was collected
• state: two-letter U.S. state abbreviation
• ili: percent ILI
• version: 1 => 2015-09; 2 => 2016-06

Similarly, we received a one-time data dump from the CDC containing true state-level ILINet data. The data was received on 2016-02-17 and uploaded to the database on 2016-02-18. This data is not to be distributed outside of the DELPHI group---and maybe not even within the group---without prior discussion. The fluview_state table stores this data (similar to the fluview table):

+---------------+---------+------+-----+---------+----------------+
| Field         | Type    | Null | Key | Default | Extra          |
+---------------+---------+------+-----+---------+----------------+
| id            | int(11) | NO   | PRI | NULL    | auto_increment |
| epiweek       | int(11) | NO   | MUL | NULL    |                |
| state         | char(2) | NO   | MUL | NULL    |                |
| num_ili       | int(11) | YES  |     | NULL    |                |
| num_patients  | int(11) | YES  |     | NULL    |                |
| num_providers | int(11) | YES  |     | NULL    |                |
| ili           | double  | YES  |     | NULL    |                |
| num_age_0     | int(11) | YES  |     | NULL    |                |
| num_age_1     | int(11) | YES  |     | NULL    |                |
| num_age_2     | int(11) | YES  |     | NULL    |                |
| num_age_3     | int(11) | YES  |     | NULL    |                |
| num_age_4     | int(11) | YES  |     | NULL    |                |
| num_age_5     | int(11) | YES  |     | NULL    |                |
+---------------+---------+------+-----+---------+----------------+

• id: unique identifier for each record
• epiweek: the epiweek during which the data was collected
• state: two-letter U.S. state abbreviation
• num_ili: the number of ILI cases (numerator)
• num_patients: the total number of patients (denominator)
• num_providers: the number of reporting healthcare providers
• ili: percent ILI
• num_age_0: number of cases in ages 0-4
• num_age_1: number of cases in ages 5-24
• num_age_2: number of cases in ages 25-64
• num_age_3: number of cases in ages 25-49
• num_age_4: number of cases in ages 50-64
• num_age_5: number of cases in ages 65+

Google stopped producing GFT after 2015w32, so the table is no longer being updated. For convenience, the data dictionary from the GFT updater is copied here: gft is the table where the data is stored.

+----------+-------------+------+-----+---------+----------------+
| Field    | Type        | Null | Key | Default | Extra          |
+----------+-------------+------+-----+---------+----------------+
| id       | int(11)     | NO   | PRI | NULL    | auto_increment |
| epiweek  | int(11)     | NO   | MUL | NULL    |                |
| location | varchar(64) | NO   | MUL | NULL    |                |
| num      | int(11)     | NO   |     | NULL    |                |
+----------+-------------+------+-----+---------+----------------+

• id: unique identifier for each record
• epiweek: the epiweek during which the data was collected
• location: where the data was collected (region, state, or city)
• num: the value, roughly corresponding to ILI * 1000

Ongoing changes are visible in the git repository.

• 2017-12-03
  • added source quidel
• 2017-02-07
  • added source flusurv
• 2016-11-15
  • support version for data from ilinet_state
• 2016-11-12
  • remove hardcoded secrets
• 2016-04-16
  • function get_region_states instead of hardcoded arrays
  • use new cdc data from table cdc_extract
• 2016-04-09
  • filter out unreasonable twitter rows
• 2016-04-07
  • added sources cdc and sensors
• 2016-04-06
  • added source stateili
• 2016-04-02
  • census regions for source twitter
• 2016-02-18
  • include more fields from fluview in ilinet
  • include optional auth parameter for CDC-provided state-level ILI
  • properly handle (don't cast) SQL NULL in execute_query
• 2016-01-18
  • added source meta
• 2015-12-15
  • added source nowcast
• 2015-12-11
  • added source signals
• 2015-12-03
  • move passwords to $AUTH variable
  • added source ght 2015* -11-19
  • using the new forecasts table for source delphi
• 2015-10-02
  • source delphi uses the forecasts table
• 2015-09-15
  • static placeholder for source delphi
• 2015-09-14
  • storing basic analytics in table api_analytics
  • patched SQL injection vulnerability in get_nidss_dengue
  • fixed a collation problem with the nidss_dengue table (see https://stackoverflow.com/questions/1008287/illegal-mix-of-collations-mysql-error)
• 2015-09-11
  • added source ilinet
• 2015-09-04
  • in wiki, added field value (1e6 * count / total)
• 2015-08-20
  • added source nidss_dengue
  • renamed source nidss to nidss_flu
• 2015-08-12
  • fixed SQL typo for daily wiki
• 2015-08-11
  • using wiki_meta for better performance and total hits
• 2015-08-10
  • added source nidss
• 2015-08-04
  • added message for invalid auth
• 2015-07-31
  • added auth parameter for twitter
• 2015-06-24
  • fully supporting twitter dataset
  • fully supporting wiki dataset
  • several utility methods to reduce duplicated code
  • heavy refactoring, additional documentation
• 2015-06-23
  • query fluview by specific lag
  • finished GFT support
  • rearranged get_fluview parameters
• 2015-06-08
  • basic support for the GFT dataset
• 2015-06-04
  • enabled multiple values and ranges
  • optional sort field
  • more documentation
• 2015-06-01
  • changes to fluview parameter names
  • removed all authentication code (most was commented out already)
• 2014-??-??
  • original version