You can listen to my one hour long show every other Wednesday at 1800 EET (8:00 AM PST) on e99.live. My track lists for each show can be found on a dedicated page on this site. My first show as at the end of February 2026, but I recently feel as if I’m finally getting my footing.

What is “The Millennial Hangover”, you may ask? It’s a show primarily driven by sensibilities of yours truly, an aging Millennial that’s becoming less relevant. Expect to hear to lots of music from about 2005 to present day. It’s not a nostalgia trip, but rather a focus my favorite music and new things I’ve come across. Shoutout to Sweep Frequency for driving my tastes over the years (it’s in the data, you can see here).

Previously, I was a DJ at KCR, the student-run station at San Diego State, from around 2007 to 2009. Additionally, during my semester abroad at Nanyang Technological University in Singapore I was a DJ on Radio Fusion (which I believe is now defunct).

If you don’t catch my show live, you can listen to each episode on SoundCloud, as seen below:

The post New radio show: The Millennial Hangover appeared first on Dispatches by John P. Gamboa.

This was primarily a first attempt at using ClaudeCode and better hone my Python skills during off months but turned into something I’ve iterated on several times throughout March 2026.

Previously, when I created apps, they haven’t been from scratch, but rather adaptations of people’s previous work (who built it from scratch). For example, I created heat maps of my Foursquare check-ins based on other posts I’ve found on the Internet.

What it does

Check-ins FM combines my Last.FM data and my Swarm/Foursquare check-ins into a single view. Claude helped me create the app using the stlib Python library. While fairly basic, it helped me some statistical insights about these two datasets.

Check-Ins FM borrows heavily from Felix Hagemens’ LastFMStats.com app to pull general stats about past listens and then triangulates them with the time check-in times and locations from Swarm. The app accepts multiple formats of CSVs of Last.FM listening history, as well as JSON files from Spotify directly. Swarm/Foursquare data are accepted as JSON files.

Both datasets have lots of information and tell a story of where I’ve been for the past 20 years or so. I can now see what kind of music I was listening to when I was traveling, and what kind of artists I tend to listen to most in different places. 

Background

For those that don’t know, Last.FM is a tool that lets users “scrobble”, or track, their music listens across platforms. I’ve used it on a clickwheel iPod, iPhone, Spotify and my Plex music library. It hasn’t been perfect, but I’ve been an active user since 2006.

Swarm (previously Foursquare) is an app that enables users to socially “check-in” to locations to let friends know where you are and become the mayor of a place (a person that checks in most, globally). I was an active user for about 10 years until the COVID-19 pandemic. Since then, I usually use it to check into an airport or a few places in a new country. These days, many people just use the Google Maps Timeline feature to keep a history of their whereabouts or frequented locations. 

If you’ve only ever used Spotify, you can use this app as well. Spotify’s metadata is richer than Last.fm’s, so the dashboard includes additional graphs from that data. If you choose to use both (to fill in gaps like I have), it will be able to merge the two datasets together without issue and de-duplicate any overlapping information.

Geocoding Challenges

The hardest part building this was using the location data from the Swarm check-ins and accurately assigning it to real places. As I didn’t want to purchase APIs from both social media sites, I had to utilize the Latitude and Longitude data embedded into each check-in to parse the real world location. 

In my first attempt I used the Nominatim (OpenStreetMap) geocoding system to assign city locations to each Lat/Long provided. The free API has a 1 request/second bottleneck, which means someone with lots of check-ins would take a long time to process. As such, I’ve pre-seeded about 1800 locations to try to speed up processing. If they aren’t in those locations, it then relies on the free API to assign geographical information. 

I’ve also preloaded maps for cities which have higher-than-average Swarm usage. I loaded the results from a research project on Swarm check-in data to reduce API reliance. It also gave me additional data for the following cities: Bandung, Beijing, Istanbul, Jakarta, Kuwait City, Melbourne, Moscow, New York, Palembang, Petaling Jaya, Sao Paulo, Shanghai, Sydney, Tangerang, and Tokyo.

The geocoding probably took the most time to fine-tune. Given my frequent moves, I needed to create a system that defines “home” in multiple places over time. The system needed to know where you lived to know when you’re traveling. Now, you can manually set your home city (or cities, if you’ve moved), and it’ll adjust timezones and detect travel accordingly. If you don’t fill this in, there’s a built-in algorithm that tries to guess your home from your check-in patterns.

There was also a problem of false attribution. Most of my listens happen Monday through Friday during working hours. If I checked into a restaurant at lunch, the system attributed all my afternoon listens to the restaurant instead of my workplace. As such, I added a weekday lunch suppression that ignores restaurant and cafe check-ins between 10am and 4pm in your home city only. If you’re traveling and check into a café, that attribution stays, because you probably were actually listening there.

Other Caveats

Each venue type gets a set amount of time for scrobbles. Airports get four hours (you’re still in transit), coffee shops get three hours, gyms for two hours, restaurants only 1.5, and cinemas just 30 minutes (you’re not listening to music during a movie). These windows determine how long after a check-in the system will attribute your scrobbles to that location.

The geocoding itself required a 25,000-entry suburb correction table to fix places that sit in middle areas between metros and suburbs. Airports, for example, are rarely within a city’s municipal boundaries (like Brussels Airport being in the village of Zaventem). Without these corrections, entire trips showed the wrong city. As such, it now maps every city under 500K population to its nearest metro within 30 kilometers, including alternate name spellings that the geocoder occasionally returns.

Each trip gets a “Music Intensity Score”: the number of scrobbles per day, so you can see which trips had the number of listens. My road trips from Austin to Houston scored high since I spent the 3.5-hour drive listening to music via Bluetooth.

The system also calculates a travel artist “lift” ratio which compares their share of listening while traveling versus at home. That’s how I can confirm that yes, I do in fact listen to Calexico disproportionately on Southwest road trips. It’s not just a feeling, it’s in the data.

What surprised me most wasn’t any particular technical challenge, but rather how much context the combined data actually provided. I can look at a trip to Seoul in 2012 and see not just that I was there, but that I was listening to Bonobo on the subway and The National at the hotel. Twenty years of check-ins and scrobbles turns out to be a detailed travelogue. Give it a try and please share your results with me! 

For some more technical information about this project, check out the GitHub repo ReadMe: https://github.com/jpgamboa/Check-Ins-FM

The post Check-ins FM webapp now available appeared first on Dispatches by John P. Gamboa.

I found this old record in my mom’s old LP collection when I moved back to the Sacramento-area last year. This LP from the now-defunct K108 radio station caught my eye due to its interesting artwork. It’s a collection of songs about Sacramento, all in a variety of musical styles popular in 1974. There’s some funk, folk, rock and instrumentals about the area. Not all of it is good, but it’s definitely a time capsule.

I tried my best, but I was able to digitize this record into MP3s that you can download and listen to here on Google Drive.

Some silly standouts include “Sacramento Boogie (Part II)”. I found mention of this song on this old website on the history of Earth Radio (KSFM 102, 1974-1979). Don Wright, the Program Director of Earth Radio, said:

They were having this contest where you write and record a song about Sacramento and send it in to the station. If your song was chosen a winner, it went on an album of songs about Sacramento! So Michael Sheehy, weekender Burt Baumgartner and I head down to my cousin’s recording studio in the Bay Area to record the hastily composed ‘Sacramento Boogie – Part 2.’ “We got the tape to KXOA-FM just before the deadline. We submitted the tape as a band called Pye Flasher. 

From what I can tell most of the musicians were/are amateurs. The discogs profile for this album has a couple of mis-attributions to similarly-named artists.

Doing some research on the names on White Pages and Facebook showed me that a large percentage of these artists have since passed away. The few that are still alive I messaged on Facebook but haven’t heard back from yet.

The album cover is a copy of the famous Avalon Ballroom “Girl with the Green Hair” concert poster from 1966 by Stanley Mouse and Alton Kelley:

Which itself is a copy of the 1896 art nouveau advertisement for Job Rolling Papers by Alphonse Mucha:

Listen to the whole album on YouTube:

Tracklist

The post “Sacramento Home Grown” (1974) Album Digitization Project appeared first on Dispatches by John P. Gamboa.

The shopkeepers told me in recent years wealthy Chinese buyers were buying these old European-designed maps and bringing them back their homeland. They hadn’t yet developed a specialty for these in the shop as they primarily focused on maps of New York and the East Coast. 

After combing through the collection a while I settled on a unique three panel map from the mid-1700s. It would be the oldest map in my collection. While not particularly rare, the idea of owning a map older than the country I’m from excited me more than enough to spent the $100 on a small, near A4 size page block print map. The map in question was pulled from from an edition of Abbé Prévost d’Exiles “Histoire générale des voyages” (1747-1767).  The map can be found on page 457 of the Seventh Volume in a section about Shaanxi Province. The map was created by French geographer JN Bellin and engraved by Dutch draughtsman Jacobus van der Schley. 

Most published maps at this time were not designed by people that had on-the-ground experience, but rather compiled from data, sketches and surveys from Jesuits living in China. In fact, Bellin himself created more than 50 maps from around the world for this multi-volume book.

However, upon doing some reverse image searches in Google, it appears that these maps were heavily borrowed from Jean-Baptiste Du Halde’s 1736 book Description de la Chine (published in English as “The General History of China”). Many of the maps in this book were created by cartographer-to-the-king, Jean-Baptiste Bourguignon d’Anville. Du Halde was a French Jesuit historian that specialized in China (despite never visiting himself) and a major reference point for the time period. Du Halde and d’Anville’s work are cited in the book, but the maps are near-exact copies. Here are the maps from his book for comparison:

What drew my attention to this map in particular was the peculiar use of blank space. While sea monsters and sirens often filled the vastness of the ocean on maps, Bellin’s background as a hydrographer lent a less-fanciful perspective to the open sea. Many of his maps have lots of blank space that give it a perceptively modern touch. Courage in the face of horror vacui drew my attention to the white space of the city walls and fortifications. 

After bringing the maps back to Austin from New York and getting it framed I had to put some effort into determining all of the locations displayed. While there was a historical system used by missionaries to romanize Chinese place names, they don’t often translate well to modern pinyin. The top two maps were easy enough to discern given their historical importance, but the third (bottom) map required the most work. Most of the my research laid dormant in a Google Chrome bookmark folder for more than half a decade.

Lanzhou 兰州 (Top Left) 

Romanized as “Lan-Tcheou” or “”Lan-Chew”, Lanzhou is described on the map as part of Chensi (Shaanxi) province and displays the old fortified city walls with lots of empty space. An additional description in French describes the city as “…where the Second Viceroy of Chensi lives” (ou là où vit le deuxième vice-roi de Chensi). 

While I never visited Lanzhou during my time living in China, the Muslim Hui population from this area were people I interacted with frequently when I lived in Shanghai. Whether it be from Lanzhou Lamian or my roommate Peter tutoring a family English.

At the start of the Qing Dynasty (1636-1912), Lanzhou was made the capital of the newly-formed Gansu province. However, it was still under the rule of the Viceroy of Shaan-Gan (according to Wikipedia, officially the Governor-General of Shaanxi, Gansu and Other Local Areas, in Charge of Military Affairs, Food and Wages, Tea and Horses and Governor Affairs), so it makes sense that the map incorrectly shows Lanzhou as part of Shaanxi province almost 100 years after the change. 

The map’s orientation shows South-as-up as the old fortress is on the southern bank. However the Yellow River (named as Hoang Ho) is shown to be flowing left-to-right, which would be incorrectly East-to-West. 

There isn’t much to the old Silk Road city on this map, but I find the detailed rendering of the surrounding mountains to be a counterbalance to the empty city map. While the mountain the city is named after isn’t prominent (GaoLanShan), I would say that it gives the impression of its 5,000 feet (1,500 meter) urban elevation on the outer edges of the Tibetan Plateau.

Most Silk Road maps help me mentally connect my maps of China to my modern Mongolia map I purchased in 2010. It’s easy to imagine people stopping here on bactrian camels before heading to further East within China or up toward Karakorum.

Prevost describes in the book (via Google Translate), “Although Lan-cheu is only a city of fertile rank, and depends on Kyang-hyang-fu, it is the capital of Iton and the governor’s residence, because being near the Great Wall and the principal gates of the West, one can easily send from there the necessary aid to the troops who defend the entrance to the Empire…Lan-cheu does not pass for a rich city. Between the strong Towns of this Province which are devoted to the defense of the Great Wall, we count Si-ning, To-pa, Ken-tan (m), Kan-cheu, Lyang-cheu, Ning-bya-wey & Yu-ling-wey. All these Places are guarded by Troops, under the command of as many General Officers; but the Generaliffime(?) is the one who lodges in Kan-cheu, a confidable Town, like that of So-cheu.”

Ganzhou 甘州 (Top Right)

Some 310 miles (500 kilometers) to the southeast of Lanzhou is the prefecture city of Zhangye. This Gansu province city was known by two names over hundreds of years, Ganzhou and Zhangye. It wasn’t until 1992 that the Chinese government uniformly enshrined its name as Zhangye, leaving the ancient city, now a district within the city, to be called Ganzhou.

Without going too much into the administrative organization of China throughout its history, Zhangye/Gansu changed administrative organization many times. There is a long blog post from the Zhangye Municipal People’s Government (Chinese only) on the origins of the city’s names. The post mentions that for hundreds of years the government changed the name and municipal status of Ganzhou/Zhangye as many as five times over the course of 300 years (before 1000 AD). Today, the city is known for the nearby Zhangye Danxia Geopark and its colorful sandstone mountains.

As with the previous map, this one also labels Ganzhou as part of Shaanxi province (Chensi) even though by the time of publishing it was brought under Gansu province. The Kan-Chow romanization often pointed me incorrectly to the Jiangxi Province city of Ganzhou (赣州) more than a thousand miles away. The name “Gansu” originates as a portmanteau of Ganzhou and Suzhou (肃州 – not the one Jiangsu Province). From a modern perspective both romanized place names for the map are incorrect, which made finding its modern location rather difficult. 

But enough about the city’s name: this map shows Ganzhou, oriented in a North Western upwards fashion, to show the Heihe River at the top, with its ancient city walls in focus. A small canal-like creek is shown on the left-hand side, but is not viewable on satellite maps anymore (likely diverted for the irrigation and farming in that area now). Like Lanzhou, Ganzhou was part of the oasis-strewn Hexi Corridor of the  Silk Road, and was protected with an inner city wall and an outer city wall. The detail of the city is missing in the city walls, just like the others. 

Portions of the city wall (primarily from the Ming Dynasty) still exist in Zhangye and can be seen on Baidu Maps Street View. The high desert of Ganzhou isn’t accurately represented in the map, but it does portray the sense of “otherness from Europe”, which was often the goal of maps at the time. My aesthetic enjoyment of the white space and its relation to the Silk Road are more interesting than any attempt at accuracy of “place.” 

Zhenyuan 镇远县 (Bottom) 

The bottom half probably took the most work of identifying when I first purchased this print many years ago. The romanization “Tchin-Ywen-Fou” yielded zero results for many years and required me to manually follow along the Yuan River (shown as Yuen Petite Riviere) until I found a city or village that matched its cartographic representation by using Baidu Maps. A dense urban area surrounded by tall limestone mountains along the river was a dead giveaway. Nothing else like that exists in along the Yuan River. It was definitely Zhenyuan. When repeated slowly, “Tchin-Ywen” and “Zhenyuan” sound exactly alike.

How did I not find this for so long? In Du Halde/D’Anville’s map, the city was romanized as “Tchin-Yuen-Fou”, just different enough that I wasn’t able to accurately find it for a long time. It was frustrating to know I was so close (and that Google couldn’t associate the place name), and that Bellin didn’t copy Du Halde accurately enough!

Currently, Zhenyuan Old Town (the area on the map) is part of Zhenyuan County in the Qiandongnan Miao and Dong Autonomous Prefecture. With a population of 200,0000, it’s a small town by modern Chinese standards. There isn’t an obvious connection between the regions of the previous two maps. Guizhou doesn’t have a direct connection to the Silk Road, the Yuan River doesn’t connect to the Heihe or Yellow Rivers, and there aren’t any threads connecting them culturally, or in the source text — it’s also 2,100 KM or 1,300 miles from Gansu province by car.

This map clearly has the most detail of the others (as does the source from Du Halde/D’Anville), but does make it appear that Zhenyuan to be a small enclave near the mountains. While the other maps accurately portray relatively flat areas surrounded by a larger, more distant mountain range, Zhenyuan appears as if the mountains surrounding it are standing on their own on a plain. In reality, the whole region is mountainous and goes on for many miles. 

The original Du Halde/D’Anville map of Zhenyuan is a collection of three Guizhou cities, and Ganzhou and Lanzhou are part of four maps on a page showing other Shaanxi (Chensi) province cities.

The only connection are the notable non-Han populations in both provinces. The map is at the end of the encyclopedia’s section on Shaanxi province that dovetails into Sichuan province (which isn’t close to Guizhou, either). There aren’t many maps in this volume, and many of them aren’t placed in the sections discussed. In a section on Yunnan Province, two maps of Longguan and Zhengding in Hebei Province are shown. Very likely, the maps didn’t have direct relation to the places shown. 

Of course, these are nitpicks to things that don’t detract from my overall enjoyment of the maps.

To the elites, clergymen and academics that would have owned this encyclopedia many hundreds of years ago, China was a fairly abstract place. Most people only interacted porcelain and other goods imported from there. Any illumination on the people or places were novel information to understanding the wider world. The collection of these three maps on a single page succeeded in delivering information about a far away land, even if their placement doesn’t make sense to our modern eyes and sense of Geography. 

Overall, it’s exciting to own such a piece of history that provides the context and knowledge of the time. I hope I’ll be able to find more like these in the future.

The post My Maps: Three Chinese Cities – Bellin mid-1700s appeared first on Dispatches by John P. Gamboa.

I ended up writing, “All Maps in China are Wrong” and not building the app. While the technical landscape has changed significantly over the previous two-thirds decade, and mainland China (and the US) has started looking inward more, China’s maps are still wrong.

In Spring 2019, the YouTube Channel Half as Interesting published a video giving some additional context and technical information as to how China’s maps are just slightly wrong.

It’s a perfect update to increasingly complex relationship China has to the rest of the connected world. Check it out below:

The post All Maps in China Are Wrong, Pt. 2 appeared first on Dispatches by John P. Gamboa.