Overview

A principal tax collecting agency in the U.S. state was in charge of administering the state’s tax laws and collecting state taxes for nearly 40 programs, including 1.8% of Transient Lodging rent. Without adequate data about the transient lodging facilities (revenue, number of rooms, exact locations, etc.), the tax collecting authority faced a challenge in monitoring whether the Transient Lodging facilities are paying the government as per the occupancy and required standards. The Department of Revenue suspected the lodging facilities of forging the occupancy rate by altering the number of rented spaces booked (Occupancy rate is calculated as the ratio of rented or used space to the total available space). The pressing need to aggregate information for these different categories got them to Xtract.io, and they decided to implement it for their business.

The data and insights they wanted to aggregate on partnering with Xtract.io

• To get the daily occupancy rate of the lodging facilities
• To validate the occupancy of the lodging facilities throughout the state daily
• Cross-check whether the taxes paid as per the occupancy rate are the same as the bookings made across online platforms
• Monthly report creation for audit purposes

Challenges

The complexity involved in aggregating this data was multi-fold. Here’s a quick overview of the expertise they required from Xtract.io

• A fully automated approach was required to determine the lodging facilities’ occupancy rate.
• The data on occupancy was required to be aggregated from 10+ third-party websites periodically (roughly 4 times a day)
• Site-specific bots were required to crawl multiple websites.

Solution

The data experts at Xtract.io analyzed the challenges and implemented a step-by-step solution.

A custom-built platform

A TOT(Transient Occupancy Tracker) platform was built to monitor the transient lodging information from sites periodically. Mobito, a proprietary web crawler platform, crawled and aggregated this data. The changes in booking status are identified and loaded onto the TOT database against the property reference. This is done daily, four times a day.

Data aggregation from multiple websites/channels

The data extraction from multiple websites is done with the help of numerous site-specific bots. These multiple site-specific bots extract data from the respective site and populate the information in the database. Using the web change monitoring bot, the daily changes in the online sites will be monitored, and the daily occupancy can be calculated.

Multiple site-specific bots were created in Python. Each of these bots performs a specific iterative process, such as extracting all links in a website, automatically downloading the HTML pages of the extracted links, etc.

All bots are placed sequentially and in parallel, as required & connected using a workflow that defines the flow of data from one bot to the other. The data is standardized and delivered through custom-built APIs

Data delivery and reports + analysis

Once the data is collected and the present and past day details are found, the occupancy details are stored as pdf/CSV/HTML. This is done four times a day. TOT system helps users configure periodic reports (daily/weekly/fortnightly/monthly/quarterly) on transient occupancy for the state/regions for delivery to the user’s secure email or to secure channels (SFTP/dropbox). TOT provides a secure login for up to 6 administrative users.

Results

Xtract.io helped the Department of Revenue identify the tax evaders by calculating the occupancy rates.

• The solution covered more than 95% of all transient lodging facilities in the State.
• Tax evaders were identified based on the variation in tax paid and occupancy data.
• 40% Increase in tax collection.