NYPL Labs @ Wiki ♡ Maps 2015

NYPL Labs @ Wiki ♡ Maps 2015

NYPL Labs @ Wiki ♡ Maps 2015
Helsinki, Finland, February 5, 2015

7aff8f547184534da3ca2e14e63a68a8?s=128

Mauricio Giraldo

February 05, 2015
Tweet

Transcript

  1. mauricio giraldo arteaga @mgiraldo nypl labs a few things we’ve

    learned so far wiki ♥ maps , helsinki, february 2015
  2. hei

  3. my name is mauricio

  4. None
  5. None
  6. NYPL Labs

  7. None
  8. None
  9. None
  10. we’ve been at it for 3+ years

  11. a few things we’ve learned so far

  12. disclaimer

  13. a few things we’ve i’ve learned so far

  14. four things

  15. practical

  16. 1

  17. start with a prototype

  18. None
  19. None
  20. None
  21. None
  22. None
  23. None
  24. None
  25. menus.nypl.org

  26. 200k dishes in 13k menus

  27. made the case for NYPL Labs

  28. it started with a prototype

  29. None
  30. None
  31. None
  32. None
  33. None
  34. None
  35. None
  36. None
  37. “wouldn’t it be cool if…?”

  38. None
  39. None
  40. None
  41. A FEW DAYS later…

  42. None
  43. None
  44. distill the task to its critical components

  45. prototypes help determine viability

  46. prototypes help communicate the idea

  47. –@mikeindustries “a prototype is worth a thousand meetings”

  48. None
  49. None
  50. “great! it’s just polish now, right?”

  51. 2

  52. “polish” takes (a lot of) time

  53. None
  54. None
  55. 3 months later…* *REAL NUMBER

  56. None
  57. None
  58. stereo.nypl.org

  59. None
  60. None
  61. None
  62. still, three months is not that long when you’re used

    to year-long projects
  63. …and polish is good

  64. LATER that day at 1:00AM…

  65. None
  66. None
  67. so far 47,000 images created

  68. another “smaller” example

  69. None
  70. None
  71. None
  72. None
  73. None
  74. None
  75. took about 2 weeks

  76. spacetime.nypl.org

  77. A FEW HOURS later…

  78. None
  79. a good problem to have?

  80. also, polish can wait

  81. None
  82. None
  83. could use a redesign

  84. None
  85. None
  86. None
  87. 1.1 million more dishes and that’s even though we’re not

    scanning menus as often
  88. curatingmenus.org by @katie_rawson and @trevormunoz

  89. @_BadTaste_ by @bibliotechy

  90. polish works

  91. “polish” is design

  92. …but it takes time

  93. “ok… so i just factor this in my planning”

  94. 3

  95. everything takes longer than anticipated

  96. people are bad at making estimates* *S. Grimstada, M. Jørgensena,

    2007; I. Newby-Clark, M. Ross, R. Buehler, D. Koehler, D. Griffin, 2007; V. Mahnič, T. Hovelja, 2012; and more…
  97. (yeah, even in “agile development”)

  98. people cannot estimate what they have not done

  99. –@jasonfried “planning is guessing”

  100. legacy website redesign ca. 2005

  101. None
  102. None
  103. None
  104. “let’s take 90 days to redesign this”

  105. (huge red flag)

  106. there are always surprises usually not the good kind of

    surprises
  107. None
  108. None
  109. None
  110. None
  111. “we have to have this feature”

  112. 18 months later…

  113. None
  114. None
  115. None
  116. None
  117. None
  118. None
  119. None
  120. None
  121. digitalcollections.nypl.org

  122. start with a scope or a deadline

  123. …but not both

  124. don’t take too long

  125. –@schuyler “if you're proud of your code you waited too

    long to ship it”
  126. …but take as long as it needs

  127. …and prepare for surprises

  128. 4

  129. hackathons are starting points

  130. None
  131. None
  132. year

  133. street names year

  134. use type street names year

  135. use type street names name year

  136. material use type street names name year

  137. material use type street names name class year

  138. material use type street names address name class year

  139. material use type street names address floors name class year

  140. material use type street names address floors name class year

    skylights
  141. material use type street names address floors name class year

    skylights backyards
  142. material use type street names address floors name class geo

    location year skylights backyards
  143. footprint material use type street names address floors name class

    geo location year skylights backyards
  144. footprint material use type street names address floors name class

    geo location year skylights backyards
  145. None
  146. “wouldn’t it be cool if…?”

  147. None
  148. None
  149. can we automate this?

  150. None
  151. None
  152. None
  153. None
  154. None
  155. None
  156. None
  157. None
  158. None
  159. FIVE WEEKS later…

  160. None
  161. github.com/NYPL/map-vectorizer

  162. None
  163. None
  164. ~80k building footprints in a day instead of years

  165. None
  166. this resulted in a new question

  167. “is the output any good?”

  168. …so we did a prototype

  169. None
  170. None
  171. None
  172. None
  173. buildinginspector.nypl.org

  174. 2 MONTHS later…

  175. None
  176. None
  177. None
  178. None
  179. 84%: YES 7%: FIX or about 91% good enough

  180. we ♥ our contributors

  181. we ♥ our contributors

  182. …so we did more prototypes!

  183. None
  184. None
  185. 1 million classifications so far! ~18 months since launch

  186. None
  187. None
  188. None
  189. None
  190. …and it all started in a hackathon

  191. recap

  192. None
  193. start with a prototype

  194. start with a prototype polish takes time

  195. start with a prototype polish takes time everything takes longer

    than anticipated
  196. start with a prototype polish takes time everything takes longer

    than anticipated hackathons are starting points
  197. thank you ♥ images from: NYPL digital collections - Wikimedia

    Commons - Flickr user wallyg - Giphy - John Resig - Peter Leonard mauricio giraldo arteaga @mgiraldo nypl labs