Walkably what-if sample: what one added place does to walking minutes ===================================================================== What this is. Walkably scores eight everyday needs, such as groceries and culture, for small hexagonal areas of the map, which we call cells. We took the map of two areas and scored it twice with the same rules: once as it is, once with one extra place added (a supermarket in Luxembourg, a library in the Helsinki area). Everything else is held the same, so the difference is what that one place does to walking minutes and scores. The "before" is the live map: every stored minute matches it exactly. Two cases - Luxembourg, one added supermarket (need: groceries). 31,575 cells, 170,068 residents. - Helsinki area (Helsinki, Espoo, Vantaa), one added library (need: culture). 25,729 cells, 525,140 residents. The result for the site shown on each map - Luxembourg, a site north of the city (49.7396 N, 6.1141 E) whose result sits close to the median of the ten sites below: the walk to the nearest supermarket in its own cell goes from 14.9 to 1.3 minutes. 183 cells get a shorter walk (1,282 residents; the ten-site median is 1,392); 1,045 more residents are within 10 minutes of a supermarket; 130 cells gain 1 score point or more (any cell, scores before rounding); the average score of the whole area moves from 55.66 to 55.69. - Helsinki area, a site near Pitäjänmäki (60.2217 N, 24.870 E): the walk to the nearest culture place in its own cell goes from 18.6 to 14.5 minutes. The walk follows streets: the cell next to it reaches the new library in 0.8 minutes, while the site's own cell has a longer street route, so its gain is smaller than its neighbours'. 215 cells get a shorter walk (5,066 residents); 1,848 more residents are within 10 minutes of one; 23 cells gain 1 score point or more (any cell, scores before rounding); the average score of the whole area moves from 64.00 to 64.01. This site is a second example, not the strongest one, and its 5,066 residents sit below the ten-site median of 6,437. - At both sites no cell gets a longer walk or a lower score. Ten sites: what a typical place does For each city we ran ten sites, chosen by one rule written before any result. A site is a cell on a street, at least 2 km inside the area, whose nearest place of the need is 12 minutes or more away. Cells with the most residents are taken first, and each site is at least 1.8 km from the ones taken before. The median of the ten is the fair figure for "a place like this"; the best site is not. - Luxembourg supermarket, median of ten: 1,392 residents with a shorter walk (lowest 278, highest 3,504); 976 more residents within 10 minutes (179 to 2,668). - Helsinki library, median of ten: 6,437 residents with a shorter walk (1,341 to 13,079); 3,933 more residents within 10 minutes (971 to 6,247). - In all twenty sites no cell gets a longer walk or a lower score, and the average score of the whole area moves by 0.1 point or less. Files - luxembourg-supermarket-before-after.png and helsinki-library-before-after.png: the map before and after. Colour is the walk in minutes to the nearest place of the need. A black ring marks a cell whose walk got shorter; the triangle is the added place. Grey cells have no place of that kind within reach (about 25 minutes). They hold about 9.5% of the residents in the Luxembourg area and about 7% in the Helsinki area. - luxembourg-supermarket-moved-cells.csv and helsinki-library-moved-cells.csv: every cell with a shorter walk at the site shown, plus any cell whose score rose by 1 point or more without a shorter walk. Columns: cell_q and cell_r (the cell's address on the Walkably grid), lat and lon (its centre), residents (people living in the cell, in whole people), minutes_before and minutes_after (walk to the nearest place of the need; "none" means no place within reach before), score_before and score_after (the cell's score, 0 to 100, shown with one decimal), walk_shorter (yes where the walk is shorter by more than 0.05 minute) and score_up_1_or_more (yes where the score rose by 1 point or more, counted on the scores before rounding). Count the yes values to get the figures above: Luxembourg 183 and 130, Helsinki 215 and 23. Subtracting the one-decimal score columns gives the same count for Luxembourg (130) but 30 cells for Helsinki, because a score that rises by 0.97 can show as 1.0 after rounding. Use the yes columns. - luxembourg-supermarket-ten-sites.csv and helsinki-library-ten-sites.csv: the ten sites, one row each. luxembourg-supermarket-ten-site-median.csv and helsinki-library-ten-site-median.csv: the median, lowest and highest over the ten. The last column of the Luxembourg file marks the site shown on the map. The Helsinki site shown is not one of the ten rows, so that column is empty there. What the numbers do not say - Ten sites picked because the nearest place is far away are not a random draw of the city. Your own site gets its own run. - It does not say a supermarket or a library would be allowed, built or used there. A library counts here because Walkably counts a library as a culture place. - One area per city is one example, and says nothing about other cities. - The minutes are walking minutes along the street network at 80 metres a minute, stored with a step of about 0.01 minute. Each cell keeps its street point from the "before" run, so a minute can change only because of the added place. - Residents are an estimate of where people live, built from the census count of each 1 km square shared out over cells by a population grid. They are rounded to whole people per cell, so a column can sum a few off the totals above. Contains map data from OpenStreetMap contributors (ODbL).