The paper is focused on open data in regions from the users' perspective and presents the perceived usability of various public data sets in the Czech Republic. There were 265 data sources assessed by respondents according the format, suggestions of further use, number of views, and usability for citizens, businesses, officers and other users. Only 14 datasets from national public agencies were already provided as open data, but none at the local level. The most frequent formats of data were DOC, HTML, PDF and XLS. The respondents came with 36 different suggestions for the further use of data sets. Currently, the citizens cannot see the difference in usability of open and non-open data unless there are particular applications available. Nationwide data sets were assessed in usability on average better (1.57) than local data sets (1.93). The usability of data was evaluated similarly across observed regions beside those provided by national institutions.
Open data is the term that echoes from local and world media with growing
frequency. Open data are complete, easily accessible, machine readable, using open
standards (e.g. CSV or XML) and published with an open license
Benefits and advantages of using open data provided by government and public
agencies (or open government data) are promising for the public administration and
for private sector (World Bank, 2014). But most public organizations have no or
limited interaction with data users and are often very selective in communication
The Czech Republic has committed to implement open government data by
joining the Open Government Partnership in 2011. Since then a number of OGD
activities have been started at different levels of the public sector as well as in
academia and other domains. Although there are some challenges that the Czech
Republic needs to face – like the missing official OGD catalogue, some public sector
bodies have already started to publish open government data
Enormous data volumes are generated on daily basis by municipalities and
regional authorities. The demand for public data in open format and the number of
relevant applications are expected to be steadily growing as an effect of the increase
in the level of digital skills and demand for information for example among
inhabitants in rural areas
There are two aims of the paper: firstly, to estimate the potential of open data sources at the national and regional level, and secondly, to evaluate user’s view of open data usability in the Czech Republic. 2
Particular research questions that were examined in the survey were stated such as: 1. The number of potential data sets identified for opening and the number of suggestions for the further use of open data was rather higher at local or regional administrations than at the national level. 2. There are differences in estimated usability related to the type of the municipality and the user group. 3. There is a significant difference in usability of open data and data that are not open. 4. There are differences in estimated usability related to the size of the municipality population.
Based on research questions further statistical processing was conducted. The methodological approach used in the paper is both of qualitative and quantitative nature and includes literature review, user interviews and descriptive analysis. The data were processed with association tables, descriptive statistics and tests to compare differences in values.
Data were collected through questionnaire form administered to citizens living in various regions. In total, there were 265 data sets assessed by respondents in terms of format, suggestions of further use, number of views, and usability for citizens, businesses, officers and other users. The perceived usability of data sets was ranked on the scale from 1 (the best) to 5 (the worst). Basic data are summarized in Table 1.
The collected data represent 52 municipalities from 9 regions (including the
capital Prague), 8 public institutions (7 schools and 1 school dormitory) and 11 state
bodies in the Czech Republic. The most frequently evaluated types of data sets were:
obligatory information published on public office website (also called as “e-desk”)
(33 %), budget information (33 %), annual report (13 %), newsletter (10 %) and
decision of representatives (10 %). In the survey focused on the openness to
information disclosure among between 395 municipalities (with up to 2,000
inhabitants) in one region of the Czech Republic in 2009
There were 16 various formats of data sets identified such as DOC, HTML, XLS, PDF, XML, CSV and other. Additional 6 various combinations of certain formats were used and some datasets were already provided in the form of a web application. Respondents came with particular ideas of further use of data sets in open format in 36 cases (13.6 %), while 14 (5.3 %) data sets were already provided according to open data principles. The remarkable finding is that all open data sets were provided by state administration bodies and none by local authorities. However more suggestions came for local data sets (26) than for national (10). The most suggestions for the further use of data were declared with files in PDF (9 suggestions), HTML (8), XLS (5), HTML combined with PDF (4) and CSV (4). The first research assumption was confirmed since the availability and the number of inputs for data sets use is prevailing at regional level.
The second question was explored by conducting a statistical analysis. Regarding
the nature of gathered data differences among mean values were examined by the
analysis of variance and several non-parametric tests. We found that data usability
evaluation does not have normal distribution and samples are of different size.
ANOVA test assumptions such as normality and symmetry of sample distribution
were not fully satisfied, so non-parametric tests were also employed, namely
Kruskal-Wallis, Leven’s test (means) and Welch’s test
H2 There is no significant difference between the local and national providers and the estimated usability (from the citizen’s perspective).
H3 There is no significant difference between regions and the estimated usability (from the citizen’s perspective).
H4 There is no significant difference between the evaluation of the data set usability and the user group.
H5 There is no significant difference in usability between data provided as open data and non-open data.
H6 There is no significant difference in usability between municipalities according to size of population.
Source: self-authored, 2015.
Basic descriptive statistics and results of hypotheses testing are presented in following tables (Table 3 – 12).
Firstly, we focused on differences between data format and perceived usability of data. Based on results in Table 4, we can conclude that there are no significant differences in the means and the type of format has no influence on the perceived usability of data. It is needed to add that only three data formats were included because the other had less than 10 evaluations that might affect reliability of results. Selected formats (DOC, HTML, PDF, XLS) are also typical for publishing data on web, however technically, PDF is considered to be one-star data according to the five-star data concept (Berners-Lee, 2010).
The interesting finding brings the frequency analysis of formats usability evaluation. Data sources were assessed with 1 or 2 in usability in particular formats: DOC (64.7 %), HTML (75.9 %), PDF (80.5 %) and XLS (75.9 %). So having data in one of these formats is likely to be perceived positively.
Many examples of successful open data applications are based on data sets provided by nationwide authorities such as police or national health agency, but we suppose that there are myriads of opportunities to exploit data from local or regional resources. The respondents spotted 216 various local data sources (see Table 5). However, none of local data sets was served in an open data format. There were some significant differences in average values (and medians according to the Welch’s test) of usability between national and local data sets (see Table 6). Thus, the hypothesis no. 2 (H2) has to be rejected.
The evaluation of the usability was positive (1 or 2) at local resources in 75.5 % of cases and at national resources in 89.8 % of cases.
Collected answers were categorized according the municipalities and regions where data were originated or related. In Table 7, there are four regions that provided at least 10 scores: whole Czech state administration (CZ), Olomoucký region (M), Plzeň region (P) and Ústí nad Labem region (U). The differences between regions were not statistically significant with 95 % probability (see Table 8). The other regions that were presented in responses, but not included in computation, were: South Bohemian region, Pardubický region, Vysočina region, Karlovarský region and Moravia-Silesian region.
The frequency of positive evaluation (1 or 2) of data sources usability across regions was distributed such as: national (89.8 %), Moravian-Silesian region (76.3 %), Plzeň region (78.4 %) and Ústí nad Labem region (81.8 %).
The usability from the perspective of citizen, business and public administration (PA) staff was examined and summarized in tables 9 and 10. However, the results were not unanimous and we cannot simply tell whether or not there are any differences in usability evaluation. From the descriptive statistics (see Table 9), we can see that the variance and mean values from business users remarkably differ. ANOVA and Levene’s tests that examine differences in means signalize that there are no significant differences, however, Kruskal-Wallis and Welch’s test indicate the opposite (see Table 10). Actually, the responses were given by citizens rather than by business people or public officers. This fact might have significant impact on data.
To answer the second research question supposing difference in data usability among regions and user groups, we can conclude that the usability was evaluated similarly across observed regions and between user groups. On the other hand, the data sets provided by national institutions were ranked with better score than regional sources (see Table 5).
The evaluation of usability according user groups ranked 1 or 2 in 78.1 % (citizen), 68.3 % (business) and 79.3 % (PA staff) of cases.
The third research question was to reveal the users’ point of view on usability of data provided in open format and data that are not open yet. The descriptive summary is presented in Table 11. There were only 11 open data sets evaluated by users against 254 non-open data sets. Some significant discrepancy in mean values was confirmed only by Levene’s p-value, but other test results including ANOVA prevented from rejecting the hypothesis about equality of mean values (see Table 12).
There was also no remarkable difference in the frequency of a positive evaluation of data usability between open data (63.6 %) and non-open data (78.7 %).
We can conclude that citizens can hardly see any difference between open and
non-open data in terms of the usability. Unless there are particular applications based
on open data much interest cannot be expected. The value of open data materializes
only upon its use
The fourth research question and resulting final hypothesis were to examine
differences among municipalities of different size. In the Table 13, we can see
comparison of testing differences in usability between municipalities (without
schools and national institutions) according to size of population. There were four
different categories of the population size: under 2,000 inhabitants, between 2,000
and 4,999, between 5,000 and 9,999, and over 10,000. The categories follow
guidelines of the Czech Statistical Office
Based on the survey, it can be concluded that data sets provided by national state agencies and authorities are more frequently in line with open data principles, while at the regional level there are still large resources that could be unleashed under open data concept. Citizens and businesses come in contact rather with local than national authorities and they could take better advantage of novel applications and services based on open data if they exist.
The main highlights of the survey findings are summarized under: • In the sample of 265 data sets, the most frequently published data were: obligatory information published on public office website (also called as “edesk”) (33 %), budget information (33 %), annual report (13 %), newsletter (10 %) and decision of representatives (10 %); • Out of 265 data sets, there were 216 data sets provided by local municipalities or institutions (81.5 %) and 49 by national bodies (18.5 %); • Nationwide data sets were assessed in usability on average better (1.57) than local data sets (1.93); • HTML (30.5 %), PDF (29.1 %), XLS (10.9 %) and DOC (6.4 %) are the most used data formats on municipality and public authority web sites. Their usability was assessed as positive (score 1 or 2) in at least 64.7 % cases; • The usability of data was evaluated similarly across observed regions beside those provided by nationwide institutions; • The usability of data was assessed from three perspectives: citizen (265 evaluations), business (60) and public administration staff (58); • The average usability was the best from citizen’s point of view (1.86) and the worst from the business point of view (2.25), but most of respondents were rather citizens than business people, which might have impacted results; • Currently, the citizens cannot see the difference in usability of open and nonopen data unless there are particular applications available; • The benefit from using data was not affected by the size of municipality population from the perspective of respondents, but those respondents did not need to be the actual residents in those municipalities.
The presented results are limited due to the scope of the survey. More details
about opinions of users coming from regions and local communities should be
investigated. Also the willingness of people to use tools based on open data and to
interact online with their local representatives should be examined thoroughly
Acknowledgments. The results and knowledge included herein have been obtained owing to support from the Internal grant agency of the Faculty of Economics and Management, Czech University of Life Sciences in Prague, grant no. 20151039 "Využití otevřených dat v agrárním sektoru" (The use of open data in agricultural sector).