Section 3.1 Work-From-Home Model

Work-from-home model predicts whether a worker is home-based-worker or not. A home-based worker does not have an out-of-home work location. The regional travel survey 2011 is used to estimate the model. The model was calibrated to the 2015 American Community Survey (ACS) 1 year share of workforce usually working from home. Table 3-1 shows the comparison between the target and model values.

Table 3-1. Work-From-Home Comparison
Target Work-From-Home 6.4%
Model Work-From-Home 6.29%

Section 3.2 Work Location Model

The work destination choice model predicts the usual work location for all workers in the population. The model uses size terms to capture the number and type of jobs available in a zone, as well as the employment-related “attractiveness” of a zone. The size terms are stratified by occupation and were developed from 2007-2011 ACS Public Use Microdata Sample (PUMS) data. Each worker was coded according to their occupation category, consistent with PECAS occupations, and their NAICS industry category, consistent with the model input employment data. Then the size terms were calculated by cross-tabulating workers by occupation and industry, and calculating for each industry the share of workers in each occupation category. These shares, shown in Table 3-2, are the size term coefficients. The size term coefficients were not updated as part of the calibration performed in 2018.

Table 3-2. Work Location Choice Size Terms
White Collar Services Health Retail And Food Blue Collar
Agriculture, Forestry, Fishing 0.3005 0.0908 0.0111 0.0171 0.5805
Mining, Oil 0.4115 0.0573 0.0000 0.0663 0.4649
Utilities 0.5500 0.0226 0.0027 0.0200 0.4047
Construction 0.2212 0.0100 0.0007 0.0125 0.7556
Manufacturing 0.4031 0.0232 0.0025 0.0676 0.5036
Wholesale trade 0.3682 0.0112 0.0011 0.3788 0.2407
Retail trade 0.2471 0.0264 0.0268 0.5602 0.1395
Transportation, Warehousing 0.4008 0.0528 0.0014 0.0202 0.5249
Information 0.6219 0.1366 0.0005 0.1225 0.1184
Finance, Insurance 0.7912 0.0125 0.0064 0.1830 0.0068
Real Estate 0.4207 0.0740 0.0004 0.3993 0.1055
Professional, Science, Technical 0.8401 0.0721 0.0174 0.0436 0.0269
Management 0.8676 0.0493 0.0042 0.0347 0.0442
Administrative, Support, Waste management 0.3514 0.4302 0.0192 0.0549 0.1444
Education 0.8199 0.0780 0.0219 0.0405 0.0397
Health care 0.3841 0.1146 0.4595 0.0216 0.0201
Arts, Entertainment, Recreation 0.1738 0.6212 0.0121 0.1503 0.0426
Accommodation, Food Service 0.1786 0.0595 0.0003 0.7264 0.0352
Other non-public administration 0.3029 0.3837 0.0128 0.0607 0.2399
Public administration 0.6251 0.2617 0.0273 0.0062 0.0796

Shadow Price Calibration

The work location model utilizes an iterative shadow pricing mechanism in order to match workers at their workplace to input employment totals. The shadow prices are written to a file and can be used in subsequent model runs to cut down computational time. The shadow price computation is done by setting the maximum number of iterations to 10 and running the work location model. The initial work location result and the final work location result at a TAZ level is plotted in Figure 3-1a and 3-1b. Figure 3-2 shows the convergence of the shadow pricing algorithm. The X axis is the shadow pricing iteration number and Y axis is the percentage of TAZs with workers that are not more than 5% different from the TAZ employment (for TAZs with at an employment of at least 100).

County-County Worker Flow

Table 3-3a and Table 3-3b show the county-to-county worker flows from the ACS data (Fratared to 2015 employment totals and 2015 workers at residence totals) and the estimated data respectively. The ACS data have been scaled to the model totals. Table 3-4c shows the differences in percentages between these two sets of data. As can be seen from these differences, the estimated data do not deviate much from the observed data. This fact is further established by visualizing this comparison as a scatterplot (Figure 3-3) - the fitted line (black line) closely follows the best fit 45 degree line (green line). The correlation coefficient for these two sets of data points is 0.993.

Absolute Terms

Table 3-3a. ACS 2006-2010 (CTPP) SCALED TO 2015 WORKER FLOWS
County Barrow Bartow Carroll Cherokee Clayton Cobb Coweta Dawson DeKalb Douglas Fayette Forsyth Fulton Gwinnett Hall Henry Newton Paulding Rockdale Spalding Walton Total
Barrow 16,321 0 0 10 51 283 0 40 769 20 0 314 1,315 11,221 1,791 0 101 0 162 0 729 33,128
Bartow 0 33,239 40 1,589 142 7,710 0 0 496 172 51 172 2,206 263 51 20 0 597 61 0 0 46,808
Carroll 0 152 39,543 51 506 1,912 1,022 0 455 3,491 162 40 2,570 142 20 61 0 435 20 10 40 50,633
Cherokee 101 2,084 172 52,646 678 29,930 71 263 3,794 435 61 3,248 23,950 3,056 162 192 30 729 81 0 0 121,684
Clayton 0 40 40 304 58,859 3,835 617 0 8,398 617 5,454 142 32,713 1,933 212 7,427 91 162 789 1,467 51 123,152
Cobb 51 3,481 1,589 7,245 6,597 232,896 334 121 17,980 6,476 809 1,558 93,019 9,643 304 698 253 4,948 385 425 61 388,872
Coweta 10 40 1,406 51 4,978 1,093 34,544 0 1,326 304 10,736 91 10,675 455 51 455 30 0 40 324 0 66,610
Dawson 0 30 0 40 0 132 0 5,090 121 0 0 2,125 971 486 931 0 0 0 0 0 0 9,926
DeKalb 425 233 486 486 11,545 14,854 243 10 172,843 860 637 1,396 128,990 29,880 850 3,177 1,366 101 3,359 809 385 372,936
Douglas 0 202 5,130 61 2,823 9,805 314 0 2,692 26,065 364 121 15,309 951 81 445 30 1,336 162 0 10 65,901
Fayette 0 20 91 30 7,356 1,093 1,801 0 1,730 152 25,934 30 11,636 344 0 1,245 10 0 81 1,568 0 53,122
Forsyth 111 51 40 1,022 557 2,722 20 2,206 3,855 101 71 52,019 28,949 9,876 3,349 162 0 20 101 121 51 105,404
Fulton 223 789 739 1,700 19,933 32,318 2,813 486 47,597 2,722 3,734 7,730 365,620 23,566 729 2,226 364 445 769 233 243 514,978
Gwinnett 2,813 263 192 536 4,938 11,565 263 293 60,326 607 374 7,032 78,135 256,654 11,555 1,123 1,386 132 2,570 526 3,299 444,584
Hall 395 20 0 81 182 395 10 1,022 1,184 51 30 2,216 2,479 7,882 71,881 61 0 20 111 0 30 88,051
Henry 0 71 0 91 22,797 1,801 283 0 6,719 192 1,538 61 15,380 1,356 0 42,153 557 20 1,052 4,412 61 98,544
Newton 51 30 0 0 1,174 385 91 30 5,768 61 40 30 4,472 1,943 111 1,983 20,217 142 9,693 405 992 47,617
Paulding 0 1,781 2,621 931 1,305 28,119 182 0 1,811 6,496 111 233 10,473 1,184 30 71 40 22,392 10 61 0 77,851
Rockdale 142 0 71 30 1,336 668 0 0 7,609 223 172 81 5,332 1,963 0 1,326 3,400 0 16,837 0 1,143 40,332
Spalding 0 10 0 0 1,285 51 324 0 172 10 820 0 789 51 10 1,629 10 20 10 18,567 0 23,758
Walton 1,234 40 304 10 293 364 20 20 3,440 172 0 71 1,548 9,714 516 354 2,054 0 2,074 51 18,416 40,696
Total 21,876 42,578 52,464 66,913 147,335 381,931 42,953 9,582 349,086 49,226 51,098 78,711 836,532 372,561 92,634 64,809 29,940 31,499 38,369 28,979 25,509 2,814,587
Table 3-3b. CT-RAMP MODEL WORKER FLOWS
Home County Barrow Bartow Carroll Cherokee Clayton Cobb Coweta Dawson DeKalb Douglas Fayette Forsyth Fulton Gwinnett Hall Henry Newton Paulding Rockdale Spalding Walton Total
Barrow 13,517 2 0 37 72 306 0 41 1,744 6 1 812 2,087 10,991 2,645 74 409 2 483 2 2,060 35,291
Bartow 0 28,528 107 3,045 70 9,968 2 14 572 381 8 300 3,033 241 21 7 0 1,118 4 0 0 47,419
Carroll 0 145 35,803 50 499 2,666 1,800 0 502 4,145 564 5 3,529 68 0 34 2 1,262 11 22 1 51,108
Cherokee 24 2,914 37 45,525 464 29,158 4 356 4,058 386 35 4,992 24,657 4,315 442 35 2 610 30 2 10 118,056
Clayton 4 7 64 60 51,278 4,348 604 0 11,548 562 4,720 72 39,462 1,564 1 5,519 178 55 567 1,041 47 121,701
Cobb 18 3,566 644 9,131 5,923 229,846 230 68 22,682 7,122 570 2,147 94,222 8,028 113 342 47 5,091 213 28 16 390,047
Coweta 0 4 2,053 11 4,123 1,800 33,095 0 1,620 1,507 9,438 5 10,498 154 0 769 8 151 61 1,188 0 66,485
Dawson 17 25 0 512 3 290 0 4,411 153 1 0 2,106 1,294 732 1,382 0 1 2 0 0 3 10,932
DeKalb 92 62 38 261 16,696 15,337 157 12 158,465 787 516 671 144,980 19,595 235 2,171 779 128 3,183 166 366 364,697
Douglas 1 196 2,881 138 2,660 10,853 916 0 3,801 23,122 886 42 19,326 624 1 209 13 1,566 81 40 4 67,360
Fayette 0 1 211 10 6,908 1,933 2,738 0 2,290 672 22,370 16 11,059 294 0 1,881 36 62 125 1,500 10 52,116
Forsyth 222 151 0 3,041 208 4,044 1 2,502 4,335 35 9 42,703 27,270 14,083 5,783 24 12 33 34 2 81 104,573
Fulton 103 281 402 2,711 25,255 33,995 1,321 190 66,099 2,848 2,456 8,183 339,310 25,965 772 911 107 402 400 180 94 511,985
Gwinnett 4,082 46 9 1,122 3,959 10,988 29 365 55,275 351 200 11,258 72,835 263,694 10,740 1,169 1,889 75 4,745 41 4,254 447,126
Hall 872 3 0 338 53 489 0 1,254 1,744 5 1 4,334 3,960 12,478 63,619 8 31 4 54 0 203 89,450
Henry 47 2 27 22 15,855 2,324 375 0 10,703 325 2,916 56 17,441 2,284 8 41,749 2,298 27 3,487 3,743 357 104,046
Newton 169 1 1 1 1,841 549 27 2 5,175 48 242 35 4,108 2,742 46 3,418 19,329 5 7,790 342 2,007 47,878
Paulding 0 2,704 2,918 1,495 1,036 23,944 309 1 2,357 6,811 314 201 11,462 722 4 65 4 21,069 31 7 3 75,457
Rockdale 65 0 4 18 2,260 953 25 1 7,169 97 212 51 6,310 3,061 35 2,090 2,215 4 15,079 209 622 40,480
Spalding 0 0 12 1 2,194 218 451 0 573 38 1,655 1 1,780 55 0 3,176 85 2 166 16,188 7 26,602
Walton 1,772 0 0 24 483 370 1 8 3,736 29 27 396 2,444 10,053 551 585 2,985 0 2,955 25 15,334 41,778
Total 21,005 38,638 45,211 67,553 141,840 384,379 42,085 9,225 364,601 49,278 47,140 78,386 841,067 381,743 86,398 64,236 30,430 31,668 39,499 24,726 25,479 2,814,587
Table 3-3c. MODEL - Target (Difference)
Home County Barrow Bartow Carroll Cherokee Clayton Cobb Coweta Dawson DeKalb Douglas Fayette Forsyth Fulton Gwinnett Hall Henry Newton Paulding Rockdale Spalding Walton Total
Barrow -2,804 2 0 27 21 23 0 1 975 -14 1 498 772 -230 854 74 308 2 321 2 1,331 2,163
Bartow 0 -4,711 67 1,456 -72 2,258 2 14 76 209 -43 128 827 -22 -30 -13 0 521 -57 0 0 611
Carroll 0 -7 -3,740 -1 -7 754 778 0 47 654 402 -35 959 -74 -20 -27 2 827 -9 12 -39 475
Cherokee -77 830 -135 -7,121 -214 -772 -67 93 264 -49 -26 1,744 707 1,259 280 -157 -28 -119 -51 2 10 -3,628
Clayton 4 -33 24 -244 -7,581 513 -13 0 3,150 -55 -734 -70 6,749 -369 -211 -1,908 87 -107 -222 -426 -4 -1,451
Cobb -33 85 -945 1,886 -674 -3,050 -104 -53 4,702 646 -239 589 1,203 -1,615 -191 -356 -206 143 -172 -397 -45 1,175
Coweta -10 -36 647 -40 -855 707 -1,449 0 294 1,203 -1,298 -86 -177 -301 -51 314 -22 151 21 864 0 -125
Dawson 17 -5 0 472 3 158 0 -679 32 1 0 -19 323 246 451 0 1 2 0 0 3 1,006
DeKalb -333 -171 -448 -225 5,151 483 -86 2 -14,378 -73 -121 -725 15,990 -10,285 -615 -1,006 -587 27 -176 -643 -19 -8,239
Douglas 1 -6 -2,249 77 -163 1,048 602 0 1,109 -2,943 522 -79 4,017 -327 -80 -236 -17 230 -81 40 -6 1,459
Fayette 0 -19 120 -20 -448 840 937 0 560 520 -3,564 -14 -577 -50 0 636 26 62 44 -68 10 -1,006
Forsyth 111 100 -40 2,019 -349 1,322 -19 296 480 -66 -62 -9,316 -1,679 4,207 2,434 -138 12 13 -67 -119 30 -831
Fulton -120 -508 -337 1,011 5,322 1,677 -1,492 -296 18,502 126 -1,278 453 -26,310 2,399 43 -1,315 -257 -43 -369 -53 -149 -2,993
Gwinnett 1,269 -217 -183 586 -979 -577 -234 72 -5,051 -256 -174 4,226 -5,300 7,040 -815 46 503 -57 2,175 -485 955 2,542
Hall 477 -17 0 257 -129 94 -10 232 560 -46 -29 2,118 1,481 4,596 -8,262 -53 31 -16 -57 0 173 1,399
Henry 47 -69 27 -69 -6,942 523 92 0 3,984 133 1,378 -5 2,061 928 8 -404 1,741 7 2,435 -669 296 5,502
Newton 118 -29 1 1 667 164 -64 -28 -593 -13 202 5 -364 799 -65 1,435 -888 -137 -1,903 -63 1,015 261
Paulding 0 923 297 564 -269 -4,175 127 1 546 315 203 -32 989 -462 -26 -6 -36 -1,323 21 -54 3 -2,394
Rockdale -77 0 -67 -12 924 285 25 1 -440 -126 40 -30 978 1,098 35 764 -1,185 4 -1,758 209 -521 148
Spalding 0 -10 12 1 909 167 127 0 401 28 835 1 991 4 -10 1,547 75 -18 156 -2,379 7 2,844
Walton 538 -40 -304 14 190 6 -19 -12 296 -143 27 325 896 339 35 231 931 0 881 -26 -3,082 1,082
Total -871 -3,940 -7,253 640 -5,495 2,448 -868 -357 15,515 52 -3,958 -325 4,535 9,182 -6,236 -573 490 169 1,130 -4,253 -30 0

Percentage Terms

Table 3-4a. ACS 2006-2010 (CTPP) SCALED TO 2015 WORKER FLOWS
County Barrow Bartow Carroll Cherokee Clayton Cobb Coweta Dawson DeKalb Douglas Fayette Forsyth Fulton Gwinnett Hall Henry Newton Paulding Rockdale Spalding Walton Total
Barrow 49% 1% 2% 1% 4% 34% 5% 2% 100%
Bartow 71% 3% 16% 1% 5% 1% 1% 100%
Carroll 78% 1% 4% 2% 1% 7% 5% 1% 100%
Cherokee 2% 43% 1% 25% 3% 3% 20% 3% 1% 100%
Clayton 48% 3% 1% 7% 1% 4% 27% 2% 6% 1% 1% 100%
Cobb 1% 2% 2% 60% 5% 2% 24% 2% 1% 100%
Coweta 2% 7% 2% 52% 2% 16% 16% 1% 1% 100%
Dawson 1% 51% 1% 21% 10% 5% 9% 100%
DeKalb 3% 4% 46% 35% 8% 1% 1% 100%
Douglas 8% 4% 15% 4% 40% 1% 23% 1% 1% 2% 100%
Fayette 14% 2% 3% 3% 49% 22% 1% 2% 3% 100%
Forsyth 1% 1% 3% 2% 4% 49% 27% 9% 3% 100%
Fulton 4% 6% 1% 9% 1% 1% 2% 71% 5% 100%
Gwinnett 1% 1% 3% 14% 2% 18% 58% 3% 1% 1% 100%
Hall 1% 1% 3% 3% 9% 82% 100%
Henry 23% 2% 7% 2% 16% 1% 43% 1% 1% 4% 100%
Newton 2% 1% 12% 9% 4% 4% 42% 20% 1% 2% 100%
Paulding 2% 3% 1% 2% 36% 2% 8% 13% 2% 29% 100%
Rockdale 3% 2% 19% 1% 13% 5% 3% 8% 42% 3% 100%
Spalding 5% 1% 1% 3% 3% 7% 78% 100%
Walton 3% 1% 1% 1% 8% 4% 24% 1% 1% 5% 5% 45% 100%
Total 1% 2% 2% 2% 5% 14% 2% 12% 2% 2% 3% 30% 13% 3% 2% 1% 1% 1% 1% 1% 100%
Table 3-4b. CT-RAMP MODEL WORKER FLOWS
Home County Barrow Bartow Carroll Cherokee Clayton Cobb Coweta Dawson DeKalb Douglas Fayette Forsyth Fulton Gwinnett Hall Henry Newton Paulding Rockdale Spalding Walton Total
Barrow 38% 1% 5% 2% 6% 31% 7% 1% 1% 6% 100%
Bartow 60% 6% 21% 1% 1% 1% 6% 1% 2% 100%
Carroll 70% 1% 5% 4% 1% 8% 1% 7% 2% 100%
Cherokee 2% 39% 25% 3% 4% 21% 4% 1% 100%
Clayton 42% 4% 9% 4% 32% 1% 5% 1% 100%
Cobb 1% 2% 2% 59% 6% 2% 1% 24% 2% 1% 100%
Coweta 3% 6% 3% 50% 2% 2% 14% 16% 1% 2% 100%
Dawson 5% 3% 40% 1% 19% 12% 7% 13% 100%
DeKalb 5% 4% 43% 40% 5% 1% 1% 100%
Douglas 4% 4% 16% 1% 6% 34% 1% 29% 1% 2% 100%
Fayette 13% 4% 5% 4% 1% 43% 21% 1% 4% 3% 100%
Forsyth 3% 4% 2% 4% 41% 26% 13% 6% 100%
Fulton 1% 5% 7% 13% 1% 2% 66% 5% 100%
Gwinnett 1% 1% 2% 12% 3% 16% 59% 2% 1% 1% 100%
Hall 1% 1% 1% 2% 5% 4% 14% 71% 100%
Henry 15% 2% 10% 3% 17% 2% 40% 2% 3% 4% 100%
Newton 4% 1% 11% 1% 9% 6% 7% 40% 16% 1% 4% 100%
Paulding 4% 4% 2% 1% 32% 3% 9% 15% 1% 28% 100%
Rockdale 6% 2% 18% 1% 16% 8% 5% 5% 37% 1% 2% 100%
Spalding 8% 1% 2% 2% 6% 7% 12% 1% 61% 100%
Walton 4% 1% 1% 9% 1% 6% 24% 1% 1% 7% 7% 37% 100%
Total 1% 1% 2% 2% 5% 14% 1% 13% 2% 2% 3% 30% 14% 3% 2% 1% 1% 1% 1% 1% 100%
Table 3-4c. MODEL - Target (Percentage Point Difference)
Home County Barrow Bartow Carroll Cherokee Clayton Cobb Coweta Dawson DeKalb Douglas Fayette Forsyth Fulton Gwinnett Hall Henry Newton Paulding Rockdale Spalding Walton Total
Barrow -11% 0.1% 0.1% 2.6% 1.4% 1.9% -2.7% 2.1% 0.2% 0.9% 0.9% 3.6%
Bartow -10.9% 0.1% 3% -0.2% 4.5% 0.1% 0.4% -0.1% 0.3% 1.7% -0.1% -0.1% 1.1% -0.1%
Carroll -8% 1.4% 1.5% 0.1% 1.2% 0.8% -0.1% 1.8% -0.1% -0.1% 1.6% -0.1%
Cherokee -0.1% 0.8% -0.1% -4.7% -0.2% 0.1% -0.1% 0.1% 0.3% 1.6% 1.2% 1.1% 0.2% -0.1% -0.1%
Clayton -0.2% -5.7% 0.5% 2.7% -0.6% -0.1% 5.9% -0.3% -0.2% -1.5% 0.1% -0.1% -0.2% -0.3%
Cobb -0.2% 0.5% -0.2% -1% 1.2% 0.2% -0.1% 0.1% 0.2% -0.4% -0.1% -0.1% -0.1%
Coweta -0.1% 1% -0.1% -1.3% 1.1% -2.1% 0.4% 1.8% -1.9% -0.1% -0.2% -0.5% -0.1% 0.5% 0.2% 1.3%
Dawson 0.2% -0.1% 4.3% 1.3% -10.9% 0.2% -2.1% 2.1% 1.8% 3.3%
DeKalb -0.1% -0.1% -0.1% 1.5% 0.2% -2.9% -0.2% 5.2% -2.6% -0.2% -0.3% -0.2% -0.2%
Douglas -3.5% 0.1% -0.3% 1.2% 0.9% 1.6% -5.2% 0.8% -0.1% 5.5% -0.5% -0.1% -0.4% 0.3% -0.1% 0.1%
Fayette 0.2% -0.6% 1.7% 1.9% 1.1% 1% -5.9% -0.7% -0.1% 1.3% 0.1% 0.1% 0.1% -0.1%
Forsyth 0.1% 0.1% 1.9% -0.3% 1.3% 0.3% 0.5% -0.1% -0.1% -8.5% -1.4% 4.1% 2.4% -0.1% -0.1% -0.1%
Fulton -0.1% -0.1% 0.2% 1.1% 0.4% -0.3% -0.1% 3.7% -0.2% 0.1% -4.7% 0.5% -0.3% -0.1%
Gwinnett 0.3% 0.1% -0.2% -0.1% -0.1% -1.2% -0.1% 0.9% -1.3% 1.2% -0.2% 0.1% 0.5% -0.1% 0.2%
Hall 0.5% 0.3% -0.1% 0.1% 0.2% 0.6% -0.1% 2.3% 1.6% 5% -10.5% -0.1% -0.1% 0.2%
Henry -0.1% -0.1% -7.9% 0.4% 0.1% 3.5% 0.1% 1.2% 1.2% 0.8% -2.7% 1.6% 2.3% -0.9% 0.3%
Newton 0.2% -0.1% 1.4% 0.3% -0.1% -0.1% -1.3% 0.4% -0.8% 1.6% -0.1% 3% -2.1% -0.3% -4.1% -0.1% 2.1%
Paulding 1.3% 0.5% 0.8% -0.3% -4.4% 0.2% 0.8% 0.7% 0.3% 1.7% -0.6% -0.8% -0.1%
Rockdale -0.2% -0.2% 2.3% 0.7% 0.1% -1.2% -0.3% 0.1% -0.1% 2.4% 2.7% 0.1% 1.9% -3% -4.5% 0.5% -1.3%
Spalding 2.8% 0.6% 0.3% 1.4% 0.1% 2.8% 3.4% 5.1% 0.3% -0.1% 0.6% -17.3%
Walton 1.2% -0.1% -0.7% 0.4% 0.5% -0.4% 0.1% 0.8% 2% 0.2% 0.1% 0.5% 2.1% 2% -0.1% -8.5%
Total -0.1% -0.3% -0.2% 0.1% 0.6% -0.1% 0.2% 0.3% -0.2% -0.2%

Scatterplot

Work Location Distance Summary

A comparison of observed and estimated distance trip frequency is shown in Figure 3-4. Only the workers who do not work-from-home and who have a work location outside of the home TAZ are included in the comparison. Table 3-5 shows the mean work location distance for different segment of workers. The table shows reasonable match between the survey targets and the models outputs.

Table 3-5. Average Target and Model Work Distance (in Miles)
Segment Target Distance Model Distance
Overall 14.24 14.08
By Person Type
Full-time worker 15.95 14.57
Part-time worker 10.13 12.20
University student 12.61 12.01
Student of driving age 7.19 12.76
By Household Income Category
Income $0 to $10k 10.88 10.56
Income $10k to $20k 10.52 11.10
Income $20k to $30k 11.89 12.36
Income $30k to $50k 13.66 12.89
Income $50k to $100k 15.14 14.72
Income gt $100k 15.11 NA
By TAZ Type
Others 8.21 10.74
SuburbRes_Exurb 13.53 15.36
Rural 18.67 19.61

Section 3.3 School Location Model

Two school location choice models are applied, one for K-12 students and one for college students. The K-12 school destination choice model predicts the usual school location for all grade-level students. As part of the calibration performed in 2018, the school enrollment data was updated with the actual numbers. This improved the school location model significantly. Only a few rounds of shadow pricing iterations were required. The distance K factors used earlier could be dropped and only a single distance squared term was used in the calibration. Table 3-6 below shows the mean distance and percentage intrazonal for different student types.

Table 3-6. Average School Distance and Percentage Intrazonal
Target
Model
Person Type Distance (in Miles) % Intrazonal Distance (in Miles) % Intrazonal
Child too young for school 4.62 8.5 3.95 10.9
Student of driving age 4.81 4.6 5.18 4.9
Student of non-driving age 4.19 11.4 3.99 11.2
University student 15.73 2.0 13.41 1.1

Shadow Price Calibration

The shadow price computation is done by setting the maximum number of iterations to 4 and running the school (and university) location model. The initial school location result and the final school location result at a TAZ level is plotted in Figure 3-5a and 3-5b. Figure 3-6 shows the convergence of the shadow pricing algorithm. The X axis is the shadow pricing iteration number and Y axis is the percentage of TAZs with students that are not more than 5% different from the TAZ enrollment.

Section 3.4 Auto Ownership Model

The auto-ownership model predicts the total number of vehicles available in a household. The ACS 2011-2015 release data was used as the benchmark. The auto-ownership model required several rounds of calibration because of the dropping of certain non-intuitive accessibility terms, and auto-ownership district constants. In addition to the calibration with respect to number of workers in the household, calibration for matching the auto ownership by household income category was also performed. Table 3-7 below shows the result after calibration by number of workers in the household.

Table 3-7. Auto Ownership Percent Share by Number of Workers
Zero Auto
One Auto
Two Auto
Three Auto
Number of Workers Target Model Target Model Target Model Target Model
0 Worker 16.3 16.0 46.8 50.1 27.8 24.7 9.0 9.2
1 Worker 4.7 5.4 48.2 49.4 34.6 33.6 12.5 11.6
2 Workers 1.8 2.2 10.6 11.7 59.4 59.7 28.1 26.4
3+ Workers 2.4 3.5 6.7 10.2 17.9 17.9 73.0 68.3
Total 6.2 6.1 34.4 34.4 39.6 39.7 19.8 19.8

The model results were compared to the observed data at a county level to establish the correctness of the spatial distribution. Table 3-8 shows the observed and the estimated shares of auto ownership level for each of the 21 counties. Based on this comparison it can be ascertained that the model is performing reasonably well at a county level.

Table 3-8. Auto Ownership Percent Share by County
Zero Auto
One Auto
Two Auto
Three Auto
County Target Model Target Model Target Model Target Model
Barrow 3.7 4.6 24.6 32.4 43.4 40.9 28.3 22.1
Bartow 4.6 5.9 29.6 33.8 40.6 40.0 25.2 20.4
Carroll 6.1 6.6 29.9 36.9 38.5 37.5 25.5 19.0
Cherokee 3.2 3.5 26.3 29.8 46.2 43.8 24.3 22.8
Clayton 7.1 5.8 43.2 36.5 32.4 36.7 17.3 21.0
Cobb 3.8 3.9 33.4 33.1 43.2 42.5 19.6 20.4
Coweta 3.7 4.8 27.1 29.8 42.4 40.2 26.8 25.1
Dawson 3.1 3.9 23.8 34.3 44.3 45.6 28.7 16.3
DeKalb 9.0 8.2 42.5 38.6 35.3 37.4 13.2 15.9
Douglas 4.3 5.1 32.6 33.1 38.5 39.8 24.5 22.0
Fayette 2.4 3.2 24.4 26.5 41.0 41.1 32.2 29.2
Forsyth 2.5 3.5 20.7 26.7 51.7 47.2 25.1 22.5
Fulton 11.7 10.8 42.2 40.9 33.7 35.3 12.4 12.9
Gwinnett 3.3 3.5 30.8 30.0 43.6 42.4 22.4 24.1
Hall 5.7 5.9 28.8 31.7 39.3 40.1 26.1 22.3
Henry 3.1 3.5 29.8 30.9 40.1 40.7 27.0 24.9
Newton 5.1 4.9 30.0 33.9 38.1 38.7 26.8 22.6
Paulding 2.8 3.8 23.7 30.1 46.6 42.4 26.9 23.8
Rockdale 4.3 4.1 35.3 31.2 40.0 40.2 20.4 24.6
Spalding 8.1 6.7 33.4 37.6 37.7 36.1 20.8 19.5
Total 6.2 6.1 34.4 34.4 39.6 39.7 19.8 19.8

Table 3-9 and Table 3-10 shows the auto ownership model result segmented by household income. Table 3-9 is in terms of absolute number of households and Table 3-10 is in terms of percentage share for different auto ownership levels. As can be seen from these two tables, during the calibration the zero auto ownership (the most crucial ownership level) match for the lower income households is done using the absolute number of households rather than the percentage shares. The reason is that the model had fewer households in those lower income categories compared to the target.

Table 3-9: Number of Households with Different Auto Ownership Levels
Zero Auto
One Auto
Two Auto
Three Auto
Total
Household Income Target Model Target Model Target Model Target Model Target Model
Income $0 to $10k 43,774 42,071 88,230 59,151 27,296 11,255 6,843 1,620 NA NA
Income $10k to $20k 39,457 38,967 121,261 83,091 46,718 26,128 13,046 4,944 220,482 153,130
Income $20k to $30k 19,949 19,943 125,594 123,119 67,783 49,255 20,627 10,054 233,952 202,371
Income $30k to $50k 17,393 16,954 192,564 203,530 158,729 154,232 61,907 56,244 430,593 430,960
Income $50k to $100k 10,662 9,793 165,098 218,792 313,455 363,211 163,012 180,850 652,227 772,646
Income greater than $100k 4,461 4,492 46,868 57,882 241,862 256,115 169,803 174,699 462,995 493,188
Total 135,696 132,220 739,616 745,565 855,842 860,196 435,238 428,411 2,166,392 2,166,392
Table 3-10: Number of Households with Different Auto Ownership Levels (percentages)
Zero Auto
One Auto
Two Auto
Three Auto
Household Income Target Model Target Model Target Model Target Model
Income $0 to $10k 26.3% 36.9% 53.1% 51.8% 16.4% 9.9% 4.1% 1.4%
Income $10k to $20k 17.9% 25.4% 55% 54.3% 21.2% 17.1% 5.9% 3.2%
Income $20k to $30k 8.5% 9.9% 53.7% 60.8% 29% 24.3% 8.8% 5%
Income $30k to $50k 4% 3.9% 44.7% 47.2% 36.9% 35.8% 14.4% 13.1%
Income $50k to $100k 1.6% 1.3% 25.3% 28.3% 48.1% 47% 25% 23.4%
Income greater than $100k 1% 0.9% 10.1% 11.7% 52.2% 51.9% 36.7% 35.4%
Total 6.3% 6.1% 34.1% 34.4% 39.5% 39.7% 20.1% 19.8%



Atlanta Regional Commission, 2019