This blog post is part of a series of posts I’m doing outlining the experiences I’ve had in a year of teaching analytics classes to NYC employees. For more information on the courses I teach, see the Github repository of course outlines and other important information as well as the other posts in this series.
The utility of analytics training for city employees isn’t something I’ve seen questioned in the packed classes I’ve been teaching over the past year as a contract instructor for the City of New York. Students from all levels in city government, from assistant commissioners to junior analysts, and all levels of management in between, have actively engaged in the process of learning new and interesting approaches to analyzing data in my classes Data Analytics for Managers, Excel Tools: Summarizing Data, Introduction to Statistical Analysis, Introduction to Open-Source Mapping with QGIS, and half-day Introduction to Open Data. They understand the importance of being able to better analyze and understand data, but as the movement to bring more and better analytics to city governments builds momentum across the country, I think it’s important to share the value I’ve seen come from this work, as well as key examples from other cities around the country.
First of all, there’s the money to be saved from improving business as usual. This comes through analytics. Cincinnati’s Office of Performance and Data Analytics was able to save the city $130,000 annually in late fees to one utility, Duke Energy. Working with 18F, the General Services Administrations technology innovation team, the DoD was able to save $150 million in costs through applying better data and technology solutions to the procurement process. Going forward, DoD plans to create their own internal data services group modeled off of the US Digital Service and the UK Government Digital Service. These are just a few examples of how government agencies are able to find cost savings through better analytics. There are literally hundreds of other examples, all showing the value of more and better analytics for improving how and where government spends it’s precious few dollars to the maximum effect possible.
This isn’t to say there aren’t already analysts working in government. Cities and government agencies at all levels have been doing analytics for a long time. During World War II, analysts working for the War Department analyzed critical elements of the war to find optimizations of tactics and strategy, as well as supply and logistics that were at the heart of the Allied victory. In New York City, analysts of some sort are found in virtually every agency looking at traffic flow on city streets, provision of services to the homeless, the demographics of city employees, and other tasks that use data to answer critical problems throughout the 5 boroughs. There is a whole track of civil servants in New York City who have analysis in their job title, but for every titled staff analyst, there are easily 2-3 other city employees who are opening Excel and trying to get answers to their problems with data.
Even some of the titled and experienced “staff analysts” have minimal training in their tool of choice (usually Excel). In many cases, they’re just treading water, knowing enough to get by and accomplish the basic tasks expected of them in their job. This isn’t unique to New York City. When I worked with the federal government, I often encountered analysts at all levels, junior and senior alike, who had minimal training with the more advanced features of Excel that make it such a powerful tool. Simple things like pivot tables and macros would blow their minds and make them excited again about their work.
This brings me to another benefit of analytics training, giving people the tools to do their job better. No one would send a bridge inspector out to perform his or her duty without first ensuring they understood all the tools and techniques of their trade. Analysts, whether formally titled as such or people who perform analytics as part of their job, are often given little training and expected to perform magic on Excel spreadsheets without so much as a basic class in Excel. The value of Excel is that it can be easy to start using, but to use it efficiently and in a way that supports the regular and reliable production of actionable data analysis, users must be trained.
Training students has two benefits for the agency, or any organization. The first is people are able to more efficiently do their job. This means answers are arrived at quicker and with a higher level of accuracy because the tool and the techniques are better understood. This increases the return on the investment put into both the tools being used (often Excel), but more importantly, the person. For their salary and compensation, they are now a more productive part of the business operation.
The second is that people who know what they are doing are happier doing it. So many of my students come into class having only the vaguest idea of how Excel works. They’ve often inherited spreadsheets that someone else designed with formulas or pivot tables and don’t understand how to change it. They can’t improve something they don’t understand and have trouble effectively managing work they don’t fully comprehend. Analysts and managers alike express their frustration with the lack of data literacy and knowledge of the analytics process.
Leaving class, students are often grateful for the deeper understanding of tools they use on a daily basis. Even little things like how to split columns, to quickly copy and paste cells, and use keyboard shortcuts can save minutes if not hours of manual data cleaning time, helping them get past the boring and repetitious tasks in their job and quickly get to the analysis. I like to think this makes them better enjoy their work by helping them cut through many mundane data cleaning tasks
A third benefit of better analytics is service. The New York City Mayor’s Office of Data Analytics has some seminal examples of this that will no doubt become text-book for city analytics, such as analysis predicting the likelihood a building contains illegal conversions or the restaurants dumping cooking grease in the sewer systems. The ClaimStat reports from the New York City Comptroller’s Office collect information about legal claims against the city to help agencies make decisions that both minimize the liability for the city and improve the experience of New Yorkers by identifying key problem areas.
In New Orleans, the Office of Performance and Accountability under Oliver Wise implemented an algorithm to help clear the massive backlog of properties awaiting inspection for either demolition or foreclosure sale. What was a decision made by one person became a decision shared by 5 people aided by an algorithm that took into account all the important variables. They’ve since moved on to develop an outreach effort targeting areas of the city without smoke alarms, a critical factor in whether a family lives or dies in a residential fire, and in so doing, created Nolalytics, a cross-department effort to drive improvements across the city through data analytics.
In New York City, with 6,000 miles of roads and over a $1 billion on construction projects annually, it’s important to understand the factors that go into road decay and cost overruns on construction projects. Answering seemingly simple questions like when is it more cost-effective to replace a road bed or simply repave has important repercussions for every New Yorker who drives, bikes, or takes surface transit in the city. Managing costs on multi-million dollar projects is key to ensuring public money is effectively spent on projects that will be built on-time, safely, and efficiently.
This brings me to the perils of bad analytics, or rather not doing analytics well. In most cases, it’s not that the mechanisms of government become malicious, they become indifferent. Government doesn’t respond to the dozens of injuries at dangerous intersections in the city and so dozens more are hurt and possibly killed. Government doesn’t see the improperly painted street segment that makes drivers think they can park near a fire hydrant without getting a ticket. Government doesn’t see how early intervention in helping people avoid eviction can help reduce the burdens on the city’s homeless shelters and lets people fall into poverty who could otherwise be saved. It lets firefighters and other first responders be hurt and possibly killed in preventable building fires caused by too many people living too close together. These not only waste money, they erode the public’s trust in their government.
Ultimately, the value in teaching analytics is less about a particular technique, tool, or immediate cost savings, it’s about helping develop a culture that is more data literate. This means relying on quality analysis to meet the complex challenges of governing a modern city rather than just carrying on business as usual. Some students come to my classes admitting they are afraid of data, and they don’t all leave sharing the love I have for data or even remembering how to create a pivot table, but they all leave with a respect for what data can tell them about their office, their agency, and their city. In teaching analytics, I’m helping teach them a respect for what analytics can do when done well and applied properly.
In the following series of posts, I’m going to detail more of the particular experiences I’ve had teaching each of the 4 full-day and 1 half-day classes I produce for New York City. In the meantime, you’re welcome to look over the course outlines for these classes and feel free to leave a comment below. If you find this work interesting and have experience teaching, please reach out to me. I’m looking to increase the number of highly qualified and motivated teachers I can call on to help me do this work. If you have a relevant background and an interest in improving how cities are run through training in the use of data, then I want to hear from you. If you represent a US city interested in developing their own curriculum of analytics training, I’m happy to talk to you as well and see if there ways my curriculum can be adapted to your needs.