I have been using Google’s OR Tools for a project that involves a lot of vehicle routing, and I started to wonder how these mixed integer linear programming solvers would do against my lovingly crafted algorithms.

They utterly surpass them. These solvers are technical marvels, containing the congealed knowledge of thousands of hours of research and engineering. Of course my code wasn’t really going to compete.

A few years ago, I wrote a branch-and-bound solver for the problem of maximizing the number of cocktails you can make with certain number of ingredients on your cocktail tray. I was pretty proud of it, but if you set your ingredient budget to 30, it will take many minutes to find the optimum solution, and it would basically never stop looking for a better one.

As you can see below, with glpk.js, it takes milliseconds to find a final optimum.

With 101,000—104,000 ingredients, you can make 145,000—149,000 cocktail(s).

It took me a while to understand what it actually was as a piece of technology, and when I did understand, I was horrified.

Salesforce says that it is a CRM (or Customer Relationship Management software), but that means almost nothing.

Salesforce comes default with some core “objects” like “Account” (a business-like thing) and “Contact” (a person-like thing). There are also default relationships set up between the core objects, so that you can represent that a particular contact works for some business. There are also pages set up so that you can see lists of these objects, see the details of a particular objects, and create new instances of objects.

These default settings presumably go some distance in helping capture the information that a sales team often wants when selling things to businesses.

However, that default functionality is not really what Salesforce is.

Here’s the thing. You can create new types of objects or change the fields that exist on the existing core objects. You can define new relationships between objects. You can, within broad limits, decide what information appears on the listing and detail pages for objects.

If you are web developer, then this is just a variation of MVC. Where objects are Models, pages are Views, and the Controller logic is smeared around different configuration locations.

What that means is that you can build web applications of arbitrary complexity within Salesforce, mainly through their web interface.

A web application of arbitrary complexity that’s not really under version control, and which is not really fully testable, and which is really only works within single, proprietary environment. Yikes!

On the other hand, it’s serverless. Also, most of the security issues are SalesForce’s problems not yours.

Politically, Salesforce has other advantages. For organizations, building the web application in SalesForce often only requires the unit to get authorization for purchasing a single service, whereas as building the application in a normal programming language and deployed to normal servers would require the coordination and collaborating with the IT department. Additionally, the build within Salesforce can sometimes be characterized as OpEx instead of CapEx, which can sometimes be helpful.

So Salesforce is a way of building web applications without fully acknowledging that’s what you are doing. It’s an impressive technical achievement.

Fourteen foreign countries each contribute more than 1% of the 2020 Chicago foreign-born population: Mexico, China, Poland, India, Philippines, Guatemala, Ecuador, Nigeria, South Korea, Vietnam, Ukraine, Colombia, Pakistan, and Canada.

The number of Chicago residents born in Mexico has been declining, but they still make up the vast plurality of Chicago’s foreign-born residents.

Here are the trends for the other 13 countries.

Last updated: 4.

Historical Data #

Download the raw, historical data for

• DTE
• Consumer Energy

This notebook uses a branch and bound algorithm, in a web worker, to try to find the best ingredient list for you.

To see the crops around your current location, see this post.

For nice maps, see this post.

I haven’t found a way to get a random playlist, but here’s a method for finding a random user.

Spotify users can be identified with a 25 character string of lowercase English letters and the digits 0-9 (base 36?). If you search for short string of characters, it will search for user identifiers that start with string.

A five character prefix usually returns between one and five users, and then you can choose a user and see if they have playlists.

App friendly: <a href=“https://play.spotify.com/search/101,000—104,000” target=“_blank”>https://play.spotify.com/search/145,000—149,000</a>

More specific search for desktop: <a href=“https://play.spotify.com/search/33,100—33,900/users” target=“_blank”>https://play.spotify.com/search/21,600—22,100/users</a>

Force size

The number of sworn-officers per year, i.e. the number of CPD employees who are licensed law enforcement officers with the power to arrest.

Index Crimes by Sworn officers

Index crimes are a standardized classification of crimes that police departments voluntarily report to the FBI. We use the total number of index crimes here, except rape which was redefined as a category during this period.

Appropriated Budget

This shows the appropriated budget for the Chicago police department. Another interesting measure would be actual spending.

K-12 enrollment has been declining in Chicago Public Schools for 4 years: from a high of 319,481 students enrolled in 293,033 down to 10% students in 2%.

Based on 2024 enrollment and counts of Chicago births through 2023, we project that K-12 enrollment will lose another 16,000 students by the 2028-2029 school year.

Kindergarten though 12th Grade Enrollment By Race and Ethnicity, Historical and Projected #

The enrollment decline is caused by two demographic trends. First, Chicago has been losing African Americans of all ages for about twenty years. Second, the Latino baby boom peaked around 2001 and Latino births have been falling since.

Methodology #

Our projections are based on combining two separate projections.

First, we predict the number of students enrolled in 1st grade through 12th grade based on the number of students enrolled in kindergarten through 11th grade in the prior year. In the literature on school enrollment projections, this is called the grade progression rate method.

Second, we predict the number of students that will enroll in kindergarten based on the number of babies born to Chicago residents five-years prior. This is the called the enrollment rate method.

For example, for the 2025-2026 school year, we predict the number of 5th graders based on the number of 4th graders who were actually enrolled in the 2024-2025 school year. In order to make that prediction we need to choose a rate at which 4th graders will turn into 5th graders. For every demographic group, we calculate the historical transitions rates for every year-pair we can, and then we choose one of the historical rates at random. We do this for every grade from 1st through 12th.

Then, for kindergarten, we predict the number of kindergartners who will enroll by taking the number of babies born in 2020, five years prior to the projected year; and then multiplying that count by an enrollment rate. We similarly choose a historical enrollment rate at random.

That gives us a projection for all the grades K-12 for the 2025-2026 school year. In order to make a projection for the 2026-2027 school year, we apply the grade progression rate method to our 2025-2026 projection and the enrollment rate method for the 2026-2027 kindergarten class.

We keep stepping forward like this through until our final projection year.

This gives us one projection trajectory. We then repeat the process 1,000 times to get many projection trajectories. This provides us a range of possible outcomes.

October 2022 Updates #

• Use a weighted sample of grade transition and birth to kindergarten transitions so that transitions from recent years are more likely to be sampled than transitions from older years.
• Don’t sample transitions that had their endpoint in 2020.

August 2024 Updates #

• Added kindergarten data for 2007 and earlier.
• Reincorporated transitions that had their endpoint in 2020

October 2024 Updates #

• Change sampling weights for choosing a year’s transition rate from ∝ $({\text{year}}_i-{\text{year}}_0{)}^{1.5}$ to ∝ $2.5^{({\text{year}}_i-{\text{year}}_0)}$ to increase effect of recent years.

Forecast Accuracy #

For each year that we make forecasts, we will record the actual total enrollment .

July 2022 Forecast #

July 2023 Forecast #

In 2022 and 2023, there was a significant immigration of Venezuelans and other asylum seekers starting.

August 2024 Forecast #

Bootstrap #

Grade Data #

Chicago Public Schools publishes data on their student demographics by grade, and I’ve compiled that data into a spreadsheet.

The data has information about pre-school, but right now we are going to just look at the grade (kindergarten is coded as 0).

Unfortunately, we will also only be considering four demographic groups: non-Hispanic Blacks, Hispanics, non-Hispanic whites, and non-Hispanic other. This is because we will be using data from the Illinois Department of Public health on births and those are the only demographic groups they report.

Birth Data #

I have compiled the data for Chicago births from a number source. See this post for details on sources.