Abdallah Gazal

Abdallah Gazal

Software Engineer

Cornell University

Information Science · Class of 2028

I build software around
the people using it.

From the interface to the systems behind it.

Ithaca, NY

← Selected work

Data analysis · Python & SQL

Beyond the Crowds

Exploring what drives national park visitation.

I enjoy visiting parks to observe nature. While staying with my aunt in the Bronx, I noticed how much park activity varied between the places I walked or took the bus to. I wanted to understand what might be behind those differences.

Those were often NYC municipal parks. This analysis looks at National Park Service units nationwide—a different set of parks, connected by the same question.

01 / Over time

A changing picture of visitation

Reported recreation visits fell 27.6% in 2020. By 2025, the total was 323.0 million—2.7% below 2024.

Reported NPS recreation visits, 1979–2025

2025 · 323,014,305 recreation visits · 400 reporting codes

1979–2025 · Reporting coverage grows from 271 to 400 codes. This is not a fixed set of parks. Hover, tap, or use the year slider.
View annual data
Reported recreation visits, 1979–2025
YearVisitsReporting codes
1979205,369,795271
1980220,463,211275
1981238,592,669280
1982244,924,579290
1983243,619,396294
1984248,785,509296
1985263,441,808303
1986281,094,850304
1987287,244,998305
1988282,451,441309
1989269,399,837314
1990255,581,467316
1991267,840,999319
1992274,694,549324
1993273,120,925327
1994268,636,169327
1995269,564,307328
1996265,796,163329
1997275,236,335336
1998286,762,265342
1999287,130,879341
2000285,891,275344
2001279,873,926345
2002277,320,211349
2003266,230,715353
2004276,908,376356
2005273,488,751356
2006272,623,980359
2007275,581,547360
2008274,852,949360
2009285,579,941360
2010281,303,769363
2011278,939,536367
2012282,765,682367
2013273,630,895370
2014292,800,082370
2015307,247,252372
2016330,971,689376
2017330,882,751379
2018318,211,833379
2019327,516,619379
2020237,064,332383
2021297,115,406388
2022311,985,998389
2023325,498,646394
2024331,863,358398
2025323,014,305400

02 / Through the year

Summer leads, but weighting matters

July averages 14.1% of a unit’s annual visits, followed by June and August. Giving every unit equal weight reveals a different profile from counting every visit equally.

July · 14.1% · Equal weight per unit

2023–2025 · 392 units with complete, positive totals in all three years. Equal weighting averages each unit-year’s monthly proportions. Hover or tap a month; focus the chart and use arrow keys.
View seasonal data
Monthly share of annual visits, 2023–2025; 392 units
MonthEqual unit weightVisit-volume weight
14.01%4.09%
24.73%4.60%
37.31%7.29%
47.94%8.23%
59.94%9.91%
612.39%12.06%
714.08%12.88%
811.88%11.60%
99.78%9.78%
107.63%8.65%
115.33%5.90%
124.97%5.01%

03 / Different kinds of places

Park designation carries information

The median unit averaged 1.16 million annual visits among National Recreation Areas, compared with 792,493 among National Parks and 27,335 among National Historic Sites.

● Median Line: middle 50% of unit averages

Select a designation to see its values.

2023–2025 annual averages · Log scale: equal spacing means a tenfold increase. Current API categories with at least five units are shown; all categories are in the table. These are not official counts of every park designation.

Purpose, attractions, access, size, and counting practices differ across these places. This comparison does not show that a designation causes higher visitation.

View all designation data
Unit means, 2023–2025; exact current API categories
DesignationUnitsMedian visits25th percentile75th percentile
Parkway212,028,2499,704,87014,351,629
Memorial Parkway16,968,9706,968,9706,968,970
National Seashore102,226,613535,6022,657,405
National River11,543,0201,543,0201,543,020
National Scenic Riverways11,264,9171,264,9171,264,917
National Recreation Area161,163,888777,6404,207,007
Ecological & Historic Preserve11,022,1391,022,1391,022,139
National Lakeshore3932,652593,3061,277,996
National Scenic Riverway1853,127853,127853,127
National Park51792,493417,1942,593,331
National and State Parks1744,823744,823744,823
National River & Recreation Area2515,918387,218644,618
Memorial2398,749252,286545,213
National Military Park9385,539192,414738,048
Park6377,312305,446466,520
National Monument and Historic Shrine1369,056369,056369,056
National Battlefield Park4367,891196,351742,983
Scenic & Recreational River1361,524361,524361,524
National Park & Preserve8275,19930,702558,728
National Historical Park and Preserve1229,162229,162229,162
National Battlefield10178,264132,807279,744
National Recreational River1145,367145,367145,367
National Historical Park59112,57636,933412,013
National Reserve190,96990,96990,969
National Memorial1585,96933,136165,822
National Preserve981,05114,013724,569
National Monument7180,20435,815184,935
National Monument & Preserve363,47031,841170,424
National Scenic River244,74129,26560,218
National Historic Site7027,33512,59198,994
International Historic Site113,80713,80713,807
National Historical Park and Ecological Preserve113,01413,01413,014
Wild River1295295295
Wild & Scenic River1197197197

04 / What I learned

Useful associations, unanswered questions

A small regression benchmark tested designation and coordinates on 367 units, holding out whole units across 25 folds. Designation was more informative than coordinates alone.

Coordinates only
−0.005
Designation only
0.273
Both together
0.303

Mean R² on log-transformed mean annual visits; the training-mean baseline scored −0.016. The combined score varied across folds (SD 0.103). These are prediction checks—not original visit-count explanatory power, causal effects, or crowding forecasts.

What the counts leave out

Recreation visits are not unique visitors, and annual totals cannot tell us how crowded a trail feels on a particular day. Reporting methods and coverage change; current metadata also misses some reporting codes.

Nearby population, accessibility, acreage, and historical designation changes remain unresolved. Local entrance counts, trail use, and visitor experience would be needed before suggesting staffing or access changes.

Source: NPS Visitor Use Statistics, 2025, with current NPS API metadata. Analysis snapshot: October 2026, commit 5f5bb2f. Four negative monthly records are quarantined and 15 partial unit-years flagged; historic totals sum available valid observations.