The Prep Stage: Data Extraction Overview | HubMeta Tutorial #13 — Transcript

Full transcript

1. 0:15Welcome back. Now that we have included
2. 0:19the articles that need to be included in
3. 0:22our systematic review, it is time to
4. 0:25extract the information we need from
5. 0:28them.
6. 0:30This is where the preface stage
7. 0:33on Hub Meta comes into play. Let's take
8. 0:35a look broadly at what is included in
9. 0:39the preface stage and then I will
10. 0:41explain them one by one.
11. 0:44First of all, we have a library page,
12. 0:47which is where it lists all the papers
13. 0:50that are in data extraction stage. And
14. 0:55on the right, we have a set of filters
15. 0:58that help us navigate through this page.
16. 1:01These filters are very, very important
17. 1:05for us to find and revise and look
18. 1:09through
19. 1:10data that is already extracted from our
20. 1:13studies. So, the multiple tabs that this
21. 1:16has are
22. 1:18data like this basically says, for
23. 1:20example, show me all the papers
24. 1:24that have any a correlation table. On
25. 1:27the right, we have a set of filters to
26. 1:30basically filter the papers that we have
27. 1:33and this is really helpful for us to
28. 1:36find a specific field, a specific data,
29. 1:39or set of
30. 1:42extracted data. After extraction, we
31. 1:45will come back to this. Now, from the
32. 1:47library, when you click on one paper, it
33. 1:52opens it up into the capture page. So,
34. 1:55we don't usually go to capture directly,
35. 1:58we go from library.
36. 2:00Uh In this page, you will see the PDF on
37. 2:04the right hand, so side-by-side view.
38. 2:07And on the left, you will see all the
39. 2:09extracted data. You can add manually the
40. 2:12templates and tables. Like in this
41. 2:15project, for example, we have extracted
42. 2:17data. Let's take a look at them.
43. 2:19Uh here, I will look at the tables.
44. 2:23And in this case, we have a correlation
45. 2:25matrix
46. 2:27uh in this study here and a regression
47. 2:31a few regression tables, actually,
48. 2:34here, which are all recorded. And
49. 2:37[snorts] in the templates tab, these are
50. 2:40data
51. 2:42uh extraction templates. Like, for
52. 2:44example,
53. 2:45uh these are the different forms that we
54. 2:47have defined to extract data. I will
55. 2:49explain that in codebook.
56. 2:51So, let's say, for example, firm sample
57. 2:55uh characteristics extraction. It says
58. 2:57the number of firms, the data years, the
59. 3:01home country, um
60. 3:04you know, the industry type, you know,
61. 3:07all the different things that
62. 3:09uh we have defined
63. 3:12um
64. 3:13to be extracted from the paper
65. 3:16uh are available here.
66. 3:18So, basically, here, you can
67. 3:21uh view and edit the data that has been
68. 3:25extracted from your paper. So,
69. 3:28uh like, um
70. 3:30for example, is endogeneity uh addressed
71. 3:34in this paper or not? Yes or no. You can
72. 3:36do it manually or you can use our AI
73. 3:38extract function, which I will explain
74. 3:40in another video. And you can directly
75. 3:44uh copy and paste from the PDF on the
76. 3:46right hand or you know, add highlights
77. 3:49and notes from your PDFs to make your
78. 3:51job of extraction
79. 3:53uh easier Uh in this page.
80. 3:56And the third page inside our prep stage
81. 4:00is codebook. This is where you define
82. 4:04what has to be extracted from your uh
83. 4:08papers. We will go into details with
84. 4:11this in our next video. And finally, AI
85. 4:14extract is where you use AI to
86. 4:20uh extract all of this data
87. 4:22automatically from your included PDFs.
88. 4:27So, let's talk about setting up the
89. 4:29codebook for our systematic review
90. 4:32project.