Metrics and Features

What insights are available in the API?

InnerSpace data is available via API and published nightly.

Definitions

Metrics are numbers, single values.

Features are objects with multiple values.

The following metrics are available via API.

Scope refers to whether a metric applies to a site (typically a floor), or a zone (a small area defined by a polygon). ..

Metric or Feature NameTypeScopeDescription

occupancy_mean_room

metric

site

For a given site, the mean of the per minute occupancy counts. Only considering zones that have a corresponding zone type listed in site_config.room_zone_types

occupancy_percent_capacity_max_room

feature

site

For a given site, which zone has the largest occ % capacity mean. Only considering zones that have a corresponding zone type listed in site_config.room_zone_types

occupancy_percent_capacity_min_room

feature

site

For a given site, which zone has the smallest occ % capacity mean. Only considering zones that have a corresponding zone type listed in site_config.room_zone_types

occupancy_pos_mean_room

metric

site

For a given site, the mean of the non-zero per minute occupancy counts. Only considering zones that have a corresponding zone type listed in site_config.room_zone_types. Where

pos indicates that the feature applies/is calculated during the times when there is a person in the space, and disregards all times the space is empty

occupancy_pos_percent_capacity_max_room

feature

site

For a given site, which zone has the largest occ pos % capacity mean. Only considering zones that have a corresponding zone type listed in site_config.room_zone_types. Where

pos indicates that the feature applies/is calculated during the times when there is a person in the space, and disregards all times the space is empty

occupancy_pos_percent_capacity_min_room

feature

site

For a given site, which zone has the smallest occ pos % capacity mean. Only considering zones that have a corresponding zone type listed in site_config.room_zone_types. Where

pos indicates that the feature applies/is calculated during the times when there is a person in the space, and disregards all times the space is empty

service_rate_max

metric

zone

The maximum service rate over a given period of time. Where service rate represents the number of people served per minute. This is typically measured for a line.

service_rate_mean

metric

zone

The average service rate over a given period of time. Where service rate represents the number of people served per minute. This is typically measured for a line.

time_available_percent_room

metric

site

Using the per minute occupancy counts, the percentage of time in which a zone/site is available, only considering zones that have a corresponding zone type listed in site_config.room_zone_types

time_available_sum_room

metric

site

Using the per minute occupancy counts, the amount of time (in seconds) in which a zone/site is available, only considering that have a corresponding zone type listed in site_config.room_zone_types

utilization_percent_room

metric

site

Using the per minute occupancy counts, the percentage of time in which a zone/site is in use, only considering zones that have a corresponding zone type listed in site_config.room_zone_types

wait_time_max

metric

zone

Maximum wait time

session_duration_mean

metric

site

For a given site, mean of the session durations during the timeframe.

  • Session duration is defined as the time (in seconds) between the end and start of the session.

  • This is different from the visit duration which is the time between the end and start of a visit (a visit is zone specific, so a session is made up of multiple visits)

However, there are sessions in which a hash starts at floor A, spends some time on other floor(s), and then later revisits floor A. Thus, the session_duration is calculated as the difference in time between the revisit to A and the original visit to A. This would seem to be an inaccurate representation of the session duration for floor A. See Figure A in the Appendix (or link below) for an example.

https://www.notion.so/innerspace/Appendix-41890def10c641ae85342ed760f2cf56

wait_time_mean

metric

zone

Mean wait time

time_on_site_dist

feature

site

Distribution representing number of hashes that spent 1 hour on site, 2 hours on site, … 24 hours on site, etc.

weekly_frequency_dist

feature

site

Distribution representing number of hashes that spent 1 day on site, 2 days, …. 7 days on site, etc.

occupancy_max

metric

zone, site

Using the occupancy counts at each minute, the maximum occupancy count during the timeframe.

occupancy_max_zone

feature

site

For a given site, which zone has the largest mean occupancy. Only considering zones that have a corresponding zone type listed in site_config.zone_occupancy_types

occupancy_mean

metric

zone, site

Using the occupancy counts at each minute (per minute occupancy), the mean occupancy of the timeframe

occupancy_min_zone

feature

site

For a given site, which zone has the smallest mean occupancy. Only considering zones that have a corresponding zone type listed in site_config.zone_occupancy_types

occupancy_percent_capacity_mean

metric

zone, site

Using the occupancy counts at each minute, mean of the occupancy in the timeframe as a percentage of the target upper occupancy. Calculated as the occupancy_mean, divided by the site_config.occupancy_target_upper for the site multiplied by 100.

occupancy_pos_max_zone

feature

site

For a given site, which zone has the largest occupancy_pos_mean. Only considering zones that have a corresponding zone type listed in site_config.zone_occupancy_types. Where

pos indicates that the feature applies/is calculated during the times when there is a person in the space, and disregards all times the space is empty

occupancy_pos_min_zone

feature

site

For a given site, which zone has the smallest occupancy_pos_mean. Only considering zones that have a corresponding zone type listed in site_config.zone_occupancy_types. Where

pos indicates that the feature applies/is calculated during the times when there is a person in the space, and disregards all times the space is empty

occupancy_pos_percent_capacity_mean

metric

zone, site

Using the occupancy counts at each minute, only including minutes where occupancy > 0, mean of the occupancy in the timeframe as a percentage of the target upper occupancy. Calculated as the occupancy_pos_mean, divided by the site_config.occupancy_target_upper for the site multiplied by 100. Where

pos indicates that the feature applies/is calculated during the times when there is a person in the space, and disregards all times the space is empty

occupancy_target

metric

zone, site

Target occupancy level

occupancy_target_lower

metric

zone, site

Lower bound of the target occupancy “range”

occupancy_target_upper

metric

zone, site

Upper bound of the target occupancy “range”

occupany_pos_mean

metric

zone, site

Using the occupancy counts at each minute, but only when occupancy > 0, the mean occupancy of the timeframe. Where

pos indicates that the feature applies/is calculated during the times when there is a person in the space, and disregards all times the space is empty

people_new_count

metric

zone, site

From session summary, counts number of hashes who are in a site/zone for the first time during the timeframe. The caveat of this metric is that since we only consider session_summary up to 1 month back of the desired start date, each hash is considered new sometime during this period. This is not quite accurate as many hashes would have already been seen beyond a month ago. For example,

The start date is Jan 1, the end date is Jan 2nd. We consider session summary roughly between Dec 1st to Jan 2nd. Suppose we saw hash ABC on Jan 1st and we know the only previous time we saw it was on Nov 25. Hash ABC will be counted as ‘new’, since in the timeframe of Dec 1st to Jan 2nd, it was seen for the first time on Jan 1st.

people_returning_count

metric

zone, site

people_unique_count - people_new_count

people_unique_count

metric

zone, site

From session summary, counts distinct hashes during the timeframe.

session_count

metric

zone, site

From session summary, count of distinct session_ids during the timeframe.

time_available

metric

zone, site

Using the per minute occupancy counts, the amount of time (in seconds) in which a zone/site is available during a given timeframe.

time_available_percent

metric

zone, site

The percentage of time in which a zone/site is available. Calculated as 100 - utilization_percent

time_on_site_mean

metric

site

For a given site, mean of each hash’s time on site during the day

Q: What is the difference between this metric and session_duration_mean? A: hashes can have multiple sessions in the same day, so time on site is the total time, rather than session specific.

utilization_percent

metric

zone, site

Using the per minute occupancy counts, the percentage of time in which a zone/site is in use.

visit_count

metric

zone

Number of visits (rows in session summary) during the timeframe

visit_duration_mean

metric

zone

For a given zone, average duration of each visit during the timeframe

weekly_frequency_mean

metric

site

Average number of distinct days each hash is on site

Last updated