Creating a word embedding in Python (using gensim) from the U.S. Tax Code

Tuesday, May 12, 2020 by Matt Pickard word embedding, natural language processing, tax

Introduction I’ve run across a need to create a word embedding for legal tax jargon, specifically centered in the United States Internal Revenue code (aka, the tax code). The tax code is available in an XML format. The XML format is convenient because it allows us to extract the different sections of the tax code. A section is the basic “level” of the tax code document hierarchy (see section 7.

Continue Reading

Calling all Computer and Data Scientists! Accounting Needs You!

Monday, May 27, 2019 by Matt Pickard accounting, data science, computer science

I’ve recently attended several presentations and engaged in several conversations that leave me panning the horizon of the Accounting world asking, “Where are all the computer and data scientists?” Seriously! Where are you? Yes, the big four Accounting firms are hiring some of you. But the need is greater. It does not seem unreasonable that much of the revolutionary changes in the field will occur more organically, from smaller start-ups.

Continue Reading

From the Perspective of One of my Students

Friday, Mar 15, 2019 by Matt Pickard data analytics, practice

Jody Padar, a student in my Accounting Data and Analytics course and a practicing accountant with a niche in automation, recently wrote this amusing perspective on the learning curve of D&A and why she’s still committed.

Continue Reading

Why Data and Analytics for Accounting?

Thursday, Mar 14, 2019 by Matt Pickard data, analytics, data and analytics, accounting

I’m an Accounting Data and Analytics (D&A) Professor at Northern Illinois University. I’m also an engineer and computer scientist at heart. I love computers. I love making computers an extension of my thinking. As I seek to teach my accounting students data and analytics (D&A), I often stop and ponder why am I teaching them this?. On good days I ask that question altruistically. On bad days, I ask that question with exasperation.

Continue Reading

Spatial Joins in R

Saturday, Feb 23, 2019 by Matt Pickard spatial joins, sp, rgdal, r

Goal The goal of this post is join spatial point and area data. I demonstrate how to identify what community each Divvy station resides in. In other words, I find the community in which the latitude and longitude of the Divvy station is and append a column to indicate the community that each Divvy station resides in. See my other post about overlaying point data onto area data with ggplot.

Continue Reading

ggplot: Overlaying Points (geom_point) on Polygons (geom_poly)

Saturday, Feb 23, 2019 by Matt Pickard rgdal, r, ggplot, map, shapefile

Goal The goal of this post is to demonstrate how to overlay geographic points onto geographic polygons. Specifically, I’ll demonstrate how map the location of Chicago Divvy bike stations (points) onto the Chicago community map (polygons from a shapefile). Data Here is the data I used: The Chicago community boundaries came in a shapefile. The divvy station data contains the latitidue and longitude of all the Divvy bike stations in the Chicago area.

Continue Reading