Analysis
The Data
Four things the reconstruction can actually answer: what the two curation methods changed, what they displaced, how closely either list was followed, and how much of each metro sits inside its own city limits.
What curation changed
Both trips ran on lists — the first self-made, the second built from published guides. The share of places visited that were already on the list barely moved: 47.4% against 52.5%. What inverted was the content. Counts are direction requests tagged by interest.
Books
11 76
Museums & history
15 44
Record stores
0 24
Food
51 69
Outdoors
14 17
Horror & occult
10 6
Film
7 0
Coffee
50 31
The same argument from the other end — where the days actually went
Of the 163 places navigated to most across the 15 cities. Published guides account for the smallest share. “On the self-made list” is an upper bound and “found on the ground” a floor — saves cannot be dated, so a place saved while standing in it counts as listed.
Interest tags are keyword-matched on destination names — a rough instrument, and places with uninformative names tag as nothing. Source: Google Takeout Maps activity.
What curation displaced
The content inverted, but the behaviour changed too — and in one direction only. Searching while already on the ground fell by roughly a sixth, while the number of distinct places a day actually reached stayed near flat. The guides answered the question that searching used to answer; they did not add stops.
Searches per direction request
1.19 0.99
How much looking-around it took to get moving
Distinct places per city-day
5.4 5.2
How much ground a day actually covered
A ratio below 1 means more places were navigated to than searched for on the ground. Both figures are corpus-wide and read verbatim. Drawn from 26,779 Maps entries between 2025-09-01 and 2026-07-18. That searching fell alongside a steady place count is consistent with the list pre-answering it — but this is observational: no part of the export records why a search did or didn't happen.
How closely each list was followed
Two different questions per city: how much of the saved list was actually navigated to, and how much of what was navigated to came off the list. Directions only — map views were excluded as circular.
Trip 1 · self-curated
Interlude · no guide at the time
Trip 2 · sourced guides
The fragmentation gap
The share of each metro living inside city limits — a proxy for how a region drew its municipal boundaries, and one of the few measures that spans the whole route. 15 of 18 cities have population data. The spread is the point: the same outcome is reached by opposite mechanisms.
510,000 in the city · 6,300,000 in the metro
A nationally cited case study in municipal fragmentation: Sandy Springs, Brookhaven, Dunwoody, Milton and others incorporated out of unincorporated Fulton and DeKalb counties from 2005, largely to control local tax base and schools.
702,250 in the city · 6,300,000 in the metro
Structurally unlike the rest: the District cannot annex or be annexed and has no county government at all, and Congress retains ultimate budget authority regardless of city-proper boundaries.
1,304,000 in the city · 8,100,000 in the metro
A genuinely polycentric metro — Fort Worth, Arlington and Plano are independent city governments rather than annexed Dallas territory, a different fragmentation mechanism from Oklahoma City's aggressive annexation.
226,610 in the city · 1,349,945 in the metro
Virginia's independent-city structure makes Richmond's own boundary unusually stable compared with the annexation and fragmentation patterns elsewhere on the corridor.
212,040 in the city · 1,257,936 in the metro
196,644 in the city · 1,120,000 in the metro
The city-proper figure sits on a declining post-2020 base.
153,701 in the city · 591,978 in the metro
Jackson's population has declined meaningfully since 2020 amid the city's prolonged water-system crisis and out-migration; the figure is a starting point, not a current count.
235,684 in the city · 837,442 in the metro
656,274 in the city · 2,336,573 in the metro
Unincorporated Clark County holds much of the built-up Las Vegas Valley — including the Strip itself — outside city-proper limits, which understates the city's functional footprint against its metro.
362,000 in the city · 1,270,000 in the metro
New Orleans's population has fluctuated significantly since Hurricane Katrina (2005), leaving the current figure less settled than most cities' estimates.
897,720 in the city · 2,805,115 in the metro
One of the fastest-growing large US cities, so the true current city-proper figure is almost certainly higher than the estimate shown.
1,650,070 in the city · 4,948,203 in the metro
178,000 in the city · 393,000 in the metro
Shreveport-proper has declined from a 2010 peak of roughly 199,000.
702,767 in the city · 1,477,926 in the metro
The metric inverts here: OKC annexed aggressively rather than fragmenting, producing a city-proper land area of roughly 620 square miles — one of the largest municipal footprints in the country — and the same outcome by the opposite route.
564,559 in the city · 918,000 in the metro
Data pending
Portland, OR · Austin, TX · Philadelphia, PA
Figures are approximate estimates pending reconciliation against a single census vintage. Annotations are drawn from each city guide's own note on how its boundary came to be.
Not yet here
Prices and spend. The card-statement export has not been run, so there is no reconstructed cost figure anywhere on this site — no prices time capsule, no spend by city. Driving distances and times. Location-history tracking was off during the trips, so there are no GPS traces to measure. Both sections land here when the exports do.