1
00:00:06,560 --> 00:00:08,640
My research focuses 
on the relationship

2
00:00:08,640 --> 00:00:10,520
between lighting and active travel.

3
00:00:10,520 --> 00:00:13,960
We look particularly at certain
aspects of lighting such as illuminance

4
00:00:13,960 --> 00:00:18,560
or the distribution of light, how 
spatially uniform it is,

5
00:00:18,560 --> 00:00:21,960
and how that affects things
like the numbers of people cycling

6
00:00:21,960 --> 00:00:25,200
or how safe people feel
when they're cycling at night.

7
00:00:25,200 --> 00:00:27,600
In the research, we took 
a number of different sets of data,

8
00:00:27,600 --> 00:00:32,400
so we had data about the number of 
people cycling in Birmingham, in the UK,

9
00:00:32,400 --> 00:00:35,000
and we compared 
that with information

10
00:00:35,000 --> 00:00:38,760
about the locations of street
lighting on different streets

11
00:00:38,760 --> 00:00:43,000
and also information
about the illuminance on those streets.

12
00:00:43,000 --> 00:00:46,280
So we were able to get night time
aerial photography

13
00:00:46,280 --> 00:00:51,040
from an open source website
and use those night time

14
00:00:51,040 --> 00:00:54,880
aerial images to estimate how 
bright streets were around

15
00:00:54,880 --> 00:00:58,400
these counters that were 
counting the number of people cycling

16
00:00:58,400 --> 00:01:00,760
 and then sort of compare the 
relationship there between

17
00:01:00,760 --> 00:01:02,040
 the brightness of the streets

18
00:01:02,040 --> 00:01:05,400
and the number of people
who were actually cycling on that street.

19
00:01:05,400 --> 00:01:10,080
So we wanted to make sure we published
our data openly,

20
00:01:10,080 --> 00:01:14,520
so we made sure we published our results in an
open access journal, PLOS ONE,

21
00:01:14,520 --> 00:01:17,280
and also PLOS ONE
has an open data policy

22
00:01:17,280 --> 00:01:19,760
where it really encourages you
to publish your data openly.

23
00:01:19,760 --> 00:01:24,280
Because a lot of our data sources were openly
available anyway and we processed them,

24
00:01:24,280 --> 00:01:27,440
we were able to share links
to those open access databases

25
00:01:27,440 --> 00:01:31,720
but also include our own data as CSV files,

26
00:01:31,720 --> 00:01:34,440
which are quite interoperable

27
00:01:34,440 --> 00:01:39,720
and could be used in any kind of analytical software.

28
00:01:39,720 --> 00:01:42,840
So our own data files were published in CSV.

29
00:01:42,840 --> 00:01:48,280
We also used GIS (geographical information system)

30
00:01:48,280 --> 00:01:52,520
as part of our analysis process
for the data

31
00:01:52,520 --> 00:01:56,080
and we used something called QGIS,
which is an open access,

32
00:01:56,080 --> 00:01:59,120
open source, freely
available piece of software.

33
00:01:59,120 --> 00:02:02,960
So, all the data files
that were related to the GIS analysis

34
00:02:02,960 --> 00:02:07,200
were also interoperable and available

35
00:02:07,200 --> 00:02:11,960
to use through this kind of 
open source GIS software as well.

36
00:02:11,960 --> 00:02:15,160
We also wanted to really make it clear

37
00:02:15,160 --> 00:02:19,440
that the analysis that we did
was very transparent as well.

38
00:02:19,440 --> 00:02:23,120
So, we used software
called R to analyse our data.

39
00:02:23,120 --> 00:02:27,120
And so, we were able to publish the code
that we'd written to analyse our data,

40
00:02:27,120 --> 00:02:29,640
openly as well, on the PLOS ONE website.

41
00:02:29,640 --> 00:02:31,920
So, anyone could see
how we'd analysed our data.

42
00:02:31,920 --> 00:02:36,600
But also, so that data could, if somebody wanted to,
they could use it themselves as well.

43
00:02:36,600 --> 00:02:40,160
I think institutional support
is really, really helpful, really valuable

44
00:02:40,160 --> 00:02:44,200
because I for one, and certainly I'm
sure many of my colleagues

45
00:02:44,200 --> 00:02:46,560
probably aren't all that familiar
with the requirements

46
00:02:46,560 --> 00:02:49,680
for actually sharing your data in a FAIR way.

47
00:02:49,680 --> 00:02:55,680
Quite often, we might think that just putting
 a raw CSV file out there is enough.

48
00:02:55,680 --> 00:02:59,720
But, there are lots of guiding principles 
about how to annotate that dataset:

49
00:02:59,720 --> 00:03:02,000
the metadata that 
would need to go with it.

50
00:03:02,000 --> 00:03:04,160
There's lots of best 
practice related to that.

51
00:03:04,160 --> 00:03:07,680
I think that's where the institution
and the library can support us with that.

52
00:03:07,680 --> 00:03:10,800
One of the challenges was knowing the best

53
00:03:10,800 --> 00:03:14,120
format to use for sharing the data.

54
00:03:14,120 --> 00:03:15,920
And I think that's certainly 
something that I would

55
00:03:15,920 --> 00:03:17,760
in the future would do better.

56
00:03:17,760 --> 00:03:19,840
I think people don't realise 
the time it takes

57
00:03:19,840 --> 00:03:23,640
to get your data into that format
where it can be shared openly

58
00:03:23,640 --> 00:03:26,160
because quite often, when 
you're immersed in a project,

59
00:03:26,160 --> 00:03:28,840
you know what your data means,
but you give that to anyone else

60
00:03:28,840 --> 00:03:30,200
and they'll have no understanding.

61
00:03:30,200 --> 00:03:31,360
So it takes a lot of time to get it

62
00:03:31,360 --> 00:03:34,800
into a format that could 
be understood by anyone.

63
00:03:34,800 --> 00:03:39,480
Commitment to that time is an important
aspect of how we can do it successfully.

64
00:03:39,480 --> 00:03:41,880
In retrospect and certainly for future projects,

65
00:03:41,880 --> 00:03:47,040
it would be great to have used something
like R markdown to create the paper.

66
00:03:47,040 --> 00:03:49,920
R Markdown is a tool or a bit of software.

67
00:03:49,920 --> 00:03:53,400
It helps you provide a 
direct link between your data,

68
00:03:53,400 --> 00:03:55,880
your analysis script as well

69
00:03:55,880 --> 00:03:58,800
and the actual text within the paper.

70
00:03:58,800 --> 00:04:02,680
So you can really see
 the dynamic link between

71
00:04:02,680 --> 00:04:04,160
all three of those aspects.

72
00:04:04,160 --> 00:04:06,920
In terms of positive 
benefits from the project,

73
00:04:06,920 --> 00:04:09,760
it forces me to learn it
and that, in turn,

74
00:04:09,760 --> 00:04:13,040
helps me pass it on to students
that I'm supervising.

75
00:04:13,040 --> 00:04:16,720
So, hopefully they will, as 
they develop their research career,

76
00:04:16,720 --> 00:04:19,240
it will become the norm for them

77
00:04:19,240 --> 00:04:23,200
whereas at the moment, it's
perhaps not quite the norm just yet.

78
00:04:23,200 --> 00:04:28,360
In terms of advice to a researcher who is 
thinking of publishing their data openly, 

79
00:04:28,360 --> 00:04:29,680
or are unsure where to start,

80
00:04:29,680 --> 00:04:32,960
first of all, I certainly recommend
speaking to the guys in the library

81
00:04:32,960 --> 00:04:36,280
who really know best practice
on how to do this and could give

82
00:04:36,280 --> 00:04:39,320
some practical examples of how
and where to do this.

83
00:04:39,320 --> 00:04:41,840
Ultimately, I think it’s just taking that first step.

84
00:04:41,840 --> 00:04:44,840
Nobody's going to do it
perfectly the first time.

85
00:04:44,840 --> 00:04:46,160
But I think just making that effort

86
00:04:46,160 --> 00:04:50,640
and showing the willingness to share
your data openly is the best

87
00:04:50,640 --> 00:04:52,160
first step you can make.