Employment Cleaning-Small Buisnesses

3 posts / 0 new
Last post
ekappelman
Employment Cleaning-Small Buisnesses

Hello,

This is a question having to do with cleaning techniques for the employment data that is used as an input for travel demand models. I am currently with the Montana Department of Transportation (MDT). At MDT we use employment data from a data collection firm in order to estimate the number of employees in each of the various employment types in each TAZ. The data from this consulting firm (or any) contains many inaccuracies. One of the major inaccuracies is the inclusion of businesses that are based in individuals homes and probably don’t really generate any trips. These include people who are selling for Mary Kay or another direct sales organization.

I had the idea of dropping out all business records in the dataset that are from businesses with two employees or fewer. I believe this would help focus trip attractions away from residential areas towards commercial centers. The reason these records cannot be vetted can be seen in a two county model currently being developed. This model uses a business data set with about 19,000 records 10,000 of which are businesses with two or fewer employees, however, these businesses only account for about 12% of the total employment.

I have shown statistically that the location of large businesses is very correlated to the location of small businesses and vice versa. I believe this supports the argument for dropping the smaller businesses because the information they provide will at least be partially captured by the large businesses that remain.

I am wondering if anyone out there would like to comment on this plan. To make it clear, I am suggesting dropping all businesses with two or fewer employees and then cleaning the remaining business data using usual methods. The hope would be that this method would produce better trip generation results with less work.

Thank you for your help.

Erik Kappelman
Transportation Planner I
Montana Department of Transportation

sramming_drcog

Erik,

You don’t say whether you’re working with a four-step model or an activity-based model.

Your proposal would probably be fine for a four-step model, which only care about the zonal employment totals by sector. (Are two-employee firms more concentrated in a particular sector, such as service?)

For an ABM, you’d want to check whether firm size is an explanatory variable anywhere. The most obvious use I can think of might be a “does firm subsidize transit passes for its employees” component.

There’s be a less direct influence in ABMs reflected by the number of people who work out of their own home. Wouldn’t Mary Kay and other direct marketing individuals still make work trips from their home to demonstrate their products and make deliveries? From time to time they’d also need to make work-related trips for supplies or possibly training.

Scott Ramming, PhD PE | Senior Travel Modeler | Transportation Planning & Operations
Direct: 303-480-6711 | Fax: 303-480-6790 | Email: sramming@drcog.org
[2016DRCOG-Logo-Side-PMS.png]
1290 Broadway • Suite 100 • Denver, Colorado 80203-5606
main: 303-455-1000 • email: drcog@drcog.org • web: www.drcog.org

[cid:image006.gif@01CF9078.A46966C0]

DRCOG helps older adults and people with disabilities age successfully and remain independent.
Call 303-480-6700 or visit us on the web!

[10] [11] [12]

From: ekappelman=mt.gov@mg.tmip.org [mailto:ekappelman=mt.gov@mg.tmip.org] On Behalf Of ekappelman
Sent: Monday, August 14, 2017 11:48 AM
To: TMIP
Subject: [TMIP] Employment Cleaning-Small Buisnesses

Hello,

This is a question having to do with cleaning techniques for the employment data that is used as an input for travel demand models. I am currently with the Montana Department of Transportation (MDT). At MDT we use employment data from a data collection firm in order to estimate the number of employees in each of the various employment types in each TAZ. The data from this consulting firm (or any) contains many inaccuracies. One of the major inaccuracies is the inclusion of businesses that are based in individuals homes and probably don’t really generate any trips. These include people who are selling for Mary Kay or another direct sales organization.

I had the idea of dropping out all business records in the dataset that are from businesses with two employees or fewer. I believe this would help focus trip attractions away from residential areas towards commercial centers. The reason these records cannot be vetted can be seen in a two county model currently being developed. This model uses a business data set with about 19,000 records 10,000 of which are businesses with two or fewer employees, however, these businesses only account for about 12% of the total employment.

I have shown statistically that the location of large businesses is very correlated to the location of small businesses and vice versa. I believe this supports the argument for dropping the smaller businesses because the information they provide will at least be partially captured by the large businesses that remain.

I am wondering if anyone out there would like to comment on this plan. To make it clear, I am suggesting dropping all businesses with two or fewer employees and then cleaning the remaining business data using usual methods. The hope would be that this method would produce better trip generation results with less work.

Thank you for your help.

Erik Kappelman
Transportation Planner I
Montana Department of Transportation
--
Full post: https://tmip.org/content/employment-cleaning-small-buisnesses
Manage my subscriptions: https://tmip.org/mailinglist
Stop emails for this post: https://tmip.org/mailinglist/unsubscribe/3403

wusun

Instead of an employment data cleaning issue, you might consider this as an issue of engineering your work location choice module in travel demand model to account for trips made by workers work from home or telecommute. In an Activity-Based Model platform, this is typically done by having a work from home/telecommute component. Also, workers work from home still generate trips, including work related travel and non-mandatory shopping, eating out, and recreational trips.

Additionally, employment is typically used as size term that represents the attractiveness of a zone as a trip destination, including passenger, truck, and commercial vehicle trips. The aggregated impact of dropping small business employment records on trip attraction should be carefully evaluated.

Wu Sun

Senior Transportation Modeler
SANDAG