City Data: Big, Open and Linked

Working Paper by Mark S. Fox (University of Toronto): “Cities are moving towards policymaking based on data. They are publishing data using Open Data standards, linking data from disparate sources, allowing the crowd to update their data with Smart Phone Apps that use Open APIs, and applying “Big Data” Techniques to discover relationships that lead to greater efficiencies.
One Big City Data example is from New York City (Schönberger & Cukier, 2013). Building owners were illegally converting their buildings into rooming houses that contained 10 times the number people they were designed for. These buildings posed a number of problems, including fire hazards, drugs, crime, disease and pest infestations. There are over 900,000 properties in New York City and only 200 inspectors who received over 25,000 illegal conversion complaints per year. The challenge was to distinguish nuisance complaints from those worth investigating where current methods were resulting in only 13% of the inspections resulting in vacate orders.
New York’s Analytics team created a dataset that combined data from 19 agencies including buildings, preservation, police, fire, tax, and building permits. By combining data analysis with expertise gleaned from inspectors (e.g., buildings that recently received a building permit were less likely to be a problem as they were being well maintained), the team was able to develop a rating system for complaints. Based on their analysis of this data, they were able to rate complaints such that in 70% of their visits, inspectors issued vacate orders; a fivefold increase in efficiency…
This paper provides an introduction to the concepts that underlie Big City Data. It explains the concepts of Open, Unified, Linked and Grounded data that lie at the heart of the Semantic Web. It then builds on this by discussing Data Analytics, which includes Statistics, Pattern Recognition and Machine Learning. Finally we discuss Big Data as the extension of Data Analytics to the Cloud where massive amounts of computing power and storage are available for processing large data sets. We use city data to illustrate each.”

Analyzing the Analyzers

catAn Introspective Survey of Data Scientists and Their Work,By Harlan Harris, Sean Murphy, Marck Vaisman: “There has been intense excitement in recent years around activities labeled “data science,” “big data,” and “analytics.” However, the lack of clarity around these terms and, particularly, around the skill sets and capabilities of their practitioners has led to inefficient communication between “data scientists” and the organizations requiring their services. This lack of clarity has frequently led to missed opportunities. To address this issue, we surveyed several hundred practitioners via the Web to explore the varieties of skills, experiences, and viewpoints in the emerging data science community.

We used dimensionality reduction techniques to divide potential data scientists into five categories based on their self-ranked skill sets (Statistics, Math/Operations Research, Business, Programming, and Machine Learning/Big Data), and four categories based on their self-identification (Data Researchers, Data Businesspeople, Data Engineers, and Data Creatives). Further examining the respondents based on their division into these categories provided additional insights into the types of professional activities, educational background, and even scale of data used by different types of Data Scientists.
In this report, we combine our results with insights and data from others to provide a better understanding of the diversity of practitioners, and to argue for the value of clearer communication around roles, teams, and careers.”