Data Sandbox

Expand all | Collapse all

The wonderful world of Points of Interest (POI)

  • 1.  The wonderful world of Points of Interest (POI)

    Pitney Bowes
    Posted 06-28-2019 04:22

    Points of Interest (POI) are being applied to address so many problems it's hard to keep track of them all. I feel like I learn a new one every time I meet clients – surprising things we've never thought about. While site selection (where to put my store or other facility) is one of the stalwart applications, a client recently showed me a very sophisticated version of this using points of interest data that included public transit locations, and Spectrum's Enterprise Routing Module to measure distances between Medicare clinics and public transportation stops.  The sophistication lies in how the distances are calculated and the result allowed this firm to improve utilization of their services by offering additional transportation options only where required. This not only assured that they complied with legislation but did it in a cost-effective way.

     

    Choosing a business location is important as well. Many years ago I worked for a company based in Bradford, UK. Our highest profile client came to visit. They parked in the car park and made their way the 20 metres to the front entrance. This moment turned out to coincide with the moment the Police decided to carry out a raid on the flats opposite, and our client was caught in the middle of a lot of police and police cars. At this point my CEO turned to me and said "Andy we need to find a better location". We moved to a shiny office in Leeds.

     

    Can you relate to these examples? The below list summarises how different industries use Points of Interest (POI) data. Are you doing something similar or different, or have you ever wondered if there was a way to improve your results by adding additional reference data to it? What would you want to achieve?

     

    Financial Services firms use POI data to:

    • Manage information coming from Points of Sale, to standardize incoming data (bank branch performance is affected by the neighborhood it is in!)
    • Detect and investigate fraudulent behavior - understanding the location of businesses in a transaction reveals insights such as are they collocated, located in unusual areas, located in areas where there is a concentration of cash based businesses. Insights that may raise red flags.
    • Understand and analyze corporate structures, who owns who.
    • Manage branch locations, analyzing competition and opportunity for openings, closures and services offered.

     

    Retail firms use POI data to:

    • Manage branch locations, analyzing competition and opportunity for openings, closures and services offered.
    • Analyse market opportunity, market share and market penetration
    • Understand work placed demand? How much spend is associated with people working in particular location.

     

    Insurance firms use POI data to:

    • Locate businesses for risk analysis and make decisions based on the locations for portfolio management – is a business located in flood area? are there natural hazards that present potential risk?
    • Understand business adjacency for risk exposure - what type of businesses are co-located, do they present a risk for example chemical manufacturing, gas/petrol station, high profile public location

     

    Adtech firms (those who specialize in marketing to mobile consumers) use POI data to:

    • Analyse mobile location data
    • Develop audience profiles
    • Identify target audiences
    • Deliver timely location relevant messaging
    • Provide location context for app developers

                   

    Telco/Wireless carriers firms use POI to data to:

    • Optimize 5G and fiber network connections to prominent businesses
    • Analyse market opportunity, market share and market penetration

     

    Real Estate firms use POI data to:

    • Help sell or rent their properties (real estate is all about location, location, location as they say)
    • Analyse markets for commercial real estate development, manage valuations and identify market opportunity

     

    Public Sector organizations use POI data to:

    • Analyse fraud and local tax avoidance and evasion
    • To identify skills and learning requirements
    • To support economic development and local infrastructure investment
    • Manage assets
    • Emergency planning


    ------------------------------
    Andy Bell
    Director
    Global Data Product Management
    Pitney Bowes Software & Data
    Leeds, UK
    ------------------------------


  • 2.  RE: The wonderful world of Points of Interest (POI)

    Pitney Bowes
    Posted 07-03-2019 18:39
    I agree Andy, POI data is fantastic and has many uses. It is amazing how much we use this type of data in everyday life, however I don't think I am the only one that has found some issues with arguably the largest consumer facing service provider of it.
    https://www.wsj.com/articles/google-maps-littered-with-fake-business-listings-harming-consumers-and-competitors-11561042283
    https://www.engadget.com/2019/06/20/wsj-google-maps-is-filled-with-fake-business-listings/?guccounter=2


    ------------------------------
    Gerry Stanley
    Pitney Bowes Australia Pty Ltd
    Macquarie Park
    ------------------------------



  • 3.  RE: The wonderful world of Points of Interest (POI)

    Pitney Bowes
    Posted 07-15-2019 16:16
    Edited by Tom Gilligan 07-15-2019 16:24
    POIs are incredibly challenging - just this weekend I read about yet another chain of retail stores going out of business (Charming Charlie) -  its not easy to keep up with all the changes even if you have a lot of staff and funding...another article on CNBC lists the scary tally for chain closings in 2019 thus far.

    ------------------------------
    Tom Gilligan
    Pitney Bowes Software, Inc.
    White River Junction, VT, USA
    ------------------------------



  • 4.  RE: The wonderful world of Points of Interest (POI)

    Pitney Bowes
    Posted 07-16-2019 03:43
    You are right Tom, it is a never ending project to maintain POI. We used to have a phrase in the UK, "It's like painting the Forth Bridge". The Forth Bridge is a famous POI just north of Edinburgh and it was always being painted. The good news is they solved that problem a few years ago, if they can solve that maybe it's possible to solve keeping POI's as fresh as possible.

    ------------------------------
    Andy Bell
    Director
    Global Data Product Management
    Pitney Bowes Software & Data
    Leeds, UK
    ------------------------------



  • 5.  RE: The wonderful world of Points of Interest (POI)

    Pitney Bowes
    Posted 26 days ago

    World Points of Interest is a best of breed data package, providing the best of many datasets conflated together.

    These datasets have their own USPs and we thrive to ingest the same and provide our clients with best of both (or more) worlds.

    But there is more to it than meets the eye, when it comes to conflating 2 or more datasets from different vendors. Each vendor has a unique naming convention of various brands/businesses.

    For example, here are just a few ways how the famous retail chain 7 ELEVEN is represented by our major vendors like Dun & Bradstreet and TomTom:-

    • 7/11
    • 7 ELEVEN
    • 7-ELEVEN
    • SEVEN 11
    • 7ELEVEN
    • 7-ELEVEN
    • 7-11
    • 711

    It would be paranormal for us to expect our end-customer to filter out all required POIs of 7-ELEVEN, taking all the above occurrences into consideration. No matter which DBMS is being used, one would have to manipulate his SQL query with the usage of wild cards and the keywords in all possible forms.

    Wouldn't it be easier if there is just a single standardized occurrence of 7-ELEVEN, and to fetch the same, one would simply query out the brandname as it exists, without any wild card characters at all.

    Standardized data is one of the biggest feats to achieve, and the business-trade-name, being the most important identifier of a POI, needs to be standardized.

    The World POI team took this humongous task as a challenge and created an all new system with the help of latest technologies like Natural Language Processing, String-Matching Algorithms like Cosine & Levenshtein, pySpark, Jupiter notebook, etc.

    As a result, a new attribute was introduced in the Points Of Interest dataset, namely BRANDNAME, which introduces one of the biggest data features craved by data specialists around the globe - standardized business-trade-name for a POI.

    With successful application of Brandname Standardization in 24 major Countries and 3000 distinct Brandnames, the process is ever expanding and under continuous improvement.

     



    ------------------------------
    Surojeet Ghosh,
    Senior Data Engineer,
    Data Development
    Noida, India
    ------------------------------



  • 6.  RE: The wonderful world of Points of Interest (POI)

    Pitney Bowes
    Posted 21 days ago
    Those are great examples, thanks for providing that. I'm excited to see where the NLP takes us.

    ------------------------------
    Teri Dickinson
    Knowledge Community Shared Account
    Shelton CT
    ------------------------------