POI Popularity Time Of Day Breakdown

Field Name
poi_id 🔑
Unique identifier of a POI. Use it to recognise the point of interest or as a foreign key to link different files and/or packages together.
date 🔑
The analysis date on which the KPI is calculated.
period 🔑
The part of the week: weekend or weekdays
time_period 🔑
The moment of the day used to provide the breakdown of the Popularity KPI as defined below. Please note that time ranges are sometimes overlapping and therefore it is not possibile to calculate the "total" popularity of a given day by adding the popularity figures of each individual timeframe. This choice was made to better represent the time of days that are actually more relevant in real life even if this means not having perfectly clean cuts between the different time intervals.The hours are divided as follows in 24hr format and in the POI time zone:Early Morning: from 04 to 08Late Morning: from 09 to 12Midday: from 11 to 14 Afternoon: from 15 to 18Evening: from 18 to 22Night: from 22 to 04
(05-10) Early Morning
Proprietary percentage KPI (uncapped range) that measures the popularity of a POI.The score is calculated through the analysis of multiple factors such as the number of geo-localised reviews and social media content contained in the calculation period.


Observations indicate a consistent trend of lower traffic volumes during the weekends compared to weekdays. I am under the impression that this comparison is not made on a proportional basis, implying a direct comparison between the total of five weekdays and the total of two weekend days. Could you kindly confirm or clarify this?

Indeed, we have verified the comparison between the number of individuals transiting from Monday to Friday and the number of individuals transiting on Saturday and Sunday.

Would it be accurate to conclude that the overall popularity of the outlet is represented by the sum of its weekend and weekday traffic?

No, it is not accurate to assert that the total popularity of the outlet is simply the sum of weekend and weekday traffic, as these two quantities do not combine in a linear fashion.

Is it appropriate to deduce the total weekday popularity based on an average calculation, or should we instead evaluate it through the summation of various time windows?

Neither assumption is correct. The total popularity of the weekday is determined through a proprietary algorithm, and thus, it cannot be directly equated to either averages or sums of time windows.

I have observed that certain time windows are unavailable in our data, yet they appear to be present on online sources. Could you shed light on any reasons that might explain this discrepancy?

We do not rely exclusively on one source of information. As a result, there may be discrepancies between the time windows available on one source and those in our Data Packs, leading to potential variations in the data presented.

Providing a high-level overview of the calculation methodology employed in this context would greatly assist in conveying the information to the business stakeholders. Could you please elaborate?

The determination of popularity across different time intervals is a multifaceted process, influenced by the proportion of individuals present during each specific period, as well as the impact that the point of interest exerts on those individuals. To accurately compute these dynamics, we employ a proprietary algorithm, ensuring a comprehensive and nuanced analysis tailored to our unique parameters.