pnorman and stereo analysis of Changesets
Following the temporary interruption of Maxar Imagery Access for OSM edits, pnorman and stereo published in the last few days analysis by changesets of Maxar usage. They used OSM Changesets Data for the last 12 months to make partial analysis from of Imagery usage by country showing a concentration in some countries from Asia and Africa.
Synthetic view of changesets and by Objects edited, 2022
I wanted to react rapidly and present a synthetic view. My excuses if any inconsistancy in resuls. The data I have on hand was extracted from the OSM Planet changeset file ending in 2022 using my own version of changesetMD. Analysis for 2022 is grouped by Continent / Sub-Continent where I divide Asia and America for North and South (South-Asia + Oceania).
The tags variable in the file provides the source and imagery_used keys, both with references to imagery used as source. And you can have sometimes simultaneous references to many images (example: Maxar, Bing, Esri). This can come from JOSM which presents us with a list of all the imageries opened in our session or is simply the fact that the contributor indicates that he used more then one image provider.
What I present is also a brief analysis where I classiy sources as :
- Maxar
- Other Major Imageries (Bing, Esri, MapBox)
- Others (all other sources - hard to classify rapidly)
- Noref (no source indicated).
Previous analysis were based on the comparison of the number of Changesets by country. I do present results for both Changesets and Objects Edited (ie. no of Nodes, ways, relations edited). The variable num_changes (no. of Objects Edited) better represents the intensity of mapping. A changeset can contain only one building (5 objects edited) or thousand of objects edited.
Analysis by changesets
For 2022, the OSM contributors created 15 million edit changesets, 17.9% with a Maxar only reference, 31.2% with reference to other Major Images, and 2% with reference to both Maxar and other Major imageries. The majors represented 51.1% of changeset sources for 2022. In comparison, there were 1.3 billion objects edited, 28% with a Maxar only reference, 34.4% with reference to other Major Images, and 4.9% with reference to both Maxar and other Major imageries. The majors represented 67.3% of Objects edited sources for 2022.
Analysis by Objects edited
As previously observed, comparison by Continents shows a higher percentage of Maxar usage in some continents. Analysis by objects edited shows that Maxar only represents 55.5% in Africa, 50.6% in South-America (this excludes Mexico added to North-America) and 50.5% in South-Asia / Oceania. In North-America, other Major imagery providers are refered for 58.8% of objects edited. Would this mean that they revised more rapidly images of North-America using Maxar Data. By contrast, in Europe the Other imageries represent 55.8% of Objects edited references (probably more availibily of other public imageries).
Tables
Table 1: OpenStreetMap Changesets, 2022
| Source / Imagery references | nd | south asia / oceania | north asia | africa | europe | south america | north america | Total |
|---|---|---|---|---|---|---|---|---|
| majors | 25,317 | 789,284 | 248,436 | 387,689 | 1,498,610 | 338,265 | 1,423,775 | 4,711,376 |
| maxar | 11,178 | 927,990 | 173,475 | 673,051 | 282,135 | 559,559 | 79,201 | 2,706,589 |
| maxar and majors | 1,105 | 110,931 | 17,893 | 27,770 | 61,144 | 36,356 | 52,769 | 307,968 |
| others | 23,413 | 310,361 | 161,239 | 156,219 | 4,946,249 | 118,579 | 404,810 | 6,120,870 |
| noref | 42,588 | 168,221 | 61,945 | 45,306 | 759,435 | 63,760 | 122,422 | 1,263,677 |
| TOTAL | 103,601 | 2,306,787 | 662,988 | 1,290,035 | 7,547,573 | 1,116,519 | 2,082,977 | 15,110,480 |
| majors | 24,4 % | 34,2 % | 37,5 % | 30,1 % | 19,9 % | 30,3 % | 68,4 % | 31,2 % |
| maxar | 10,8 % | 40,2 % | 26,2 % | 52,2 % | 3,7 % | 50,1 % | 3,8 % | 17,9 % |
| maxar and majors | 1,1 % | 4,8 % | 2,7 % | 2,2 % | 0,8 % | 3,3 % | 2,5 % | 2,0 % |
| others | 22,6 % | 13,5 % | 24,3 % | 12,1 % | 65,5 % | 10,6 % | 19,4 % | 40,5 % |
| noref | 41,1 % | 7,3 % | 9,3 % | 3,5 % | 10,1 % | 5,7 % | 5,9 % | 8,4 % |
| TOTAL | 100 % | 100 % | 100 % | 100 % | 100 % | 100 % | 100 % | 100 % |
| Changesets : count of rows from Changeset file |
Table 2: OpenStreetMap Objects edited (nodes + ways + relations), 2022
| Source / Imagery references | nd | south asia / oceania | north asia | africa | europe | south america | north america | Total |
|---|---|---|---|---|---|---|---|---|
| majors | 246,933 | 55,698,666 | 41,307,960 | 64,644,325 | 109,001,935 | 27,089,932 | 143,856,608 | 441,846,359 |
| maxar | 316,645 | 97,789,257 | 47,042,080 | 126,467,184 | 31,220,099 | 49,909,048 | 7,465,196 | 360,209,509 |
| maxar and majors | 129,679 | 16,119,891 | 7,960,065 | 9,740,725 | 11,574,609 | 9,131,001 | 8,648,204 | 63,304,174 |
| others | 559,559 | 11,722,100 | 18,756,406 | 22,407,195 | 221,156,231 | 9,306,103 | 68,468,511 | 352,376,105 |
| noref | 874,072 | 12,121,020 | 7,082,760 | 4,565,216 | 23,505,369 | 3,283,832 | 16,971,905 | 68,404,174 |
| TOTAL | 2,126,888 | 193,450,934 | 122,149,271 | 227,824,645 | 396,458,243 | 98,719,916 | 245,410,424 | 1,286,140,321 |
| majors | 11,6 % | 28,8 % | 33,8 % | 28,4 % | 27,5 % | 27,4 % | 58,6 % | 34,4 % |
| maxar | 14,9 % | 50,5 % | 38,5 % | 55,5 % | 7,9 % | 50,6 % | 3,0 % | 28,0 % |
| maxar and majors | 6,1 % | 8,3 % | 6,5 % | 4,3 % | 2,9 % | 9,2 % | 3,5 % | 4,9 % |
| others | 26,3 % | 6,1 % | 15,4 % | 9,8 % | 55,8 % | 9,4 % | 27,9 % | 27,4 % |
| noref | 41,1 % | 6,3 % | 5,8 % | 2,0 % | 5,9 % | 3,3 % | 6,9 % | 5,3 % |
| TOTAL | 100 % | 100 % | 100 % | 100 % | 100 % | 100 % | 100 % | 100 % |
| Objects edited : variable num_changes from Changeset file |