Last posts

An experiment with data from the open government partnership: Ranking countries

August 19, 2014 / maestro / ogov, open data, open government, Uncategorized

The OGP is a global organization to promote Open government on more than 60 countries.

Every member can submit a annual plan to explain which will be their action for the next year. This plan is independently reviewed and their reviews are published.

Even more, some main data about accomplishment are made public in this databases (Commitments analyzed  and Accomplishments of the plan).

Let’s make a tiny experiment with these data.

First of all we have to know that the submitted plan could incorporate new actions or existing ones. Secondly these actions could be completed or not accordingly to the plan. Thirdly and very important, these action could have a moderate, minor or a transformative effect on the countries’ governance. Last but not least, these action could be relevant (or no) for an open government plan.

In all these dimensions of analysis there are some cases in which it is not clear the qualification (or data are not available). These cases should be also considered in this experiment with data.Also severla countries did not submit their reviews on time so their data are not available (USA, UK, S. Africa, Philippines, Mexico, Indonesia and Brazil)

Now let’s create a draft metric about the performance of the different countries.

my first attempt is

Plan Ambition = log (\sum 100*(Action_i(new))* 100*Action_i(impact)* 100*Action_i(specificity))
Why the log? In order to promote more detailed action plans (more actions) but without penalize those more concise. Multiplied if the action has real impact and if the action really affects governance. With these metric we could rank countries according to these graph.

Similarly the progress will be just a minor correction on above formula, and normalized by the # of actions.
Plan Ambition = log ((\sum 100*(Action_i(new))* 100*Action_i(impact)* 100*Action_i(specificity)*Action_i(progress))/\sum _i)

And the the ranking would look like this.

Of course it is just a mere experiment. Possibly unfair with some countries because of the structure of their plan, the assigned weights or because of past experience makes them harder to be ambitious in their plan formulation.
Are we mature enoguh to create such official ranking?  I strongly believe it.

Find here the raw data calculation if you want to make your own experiment.


Find here the table which assign weights for every dimension of the actions in the plan fo every country.

New or not how many weights
0 0
New 317 1
Pre-existing 445 0.5
Unclear 21 0.5
Grand Total 783
Completion how many weights
a. Not started 116 0
b. Limited 319 0.25
c. Substantial 204 0.5
d. Complete 271 1
NA 8 0
Unclear 23 0
Withdrawn 17 0
Grand Total 958
Impact how many weights
0 0
Minor 247 0.25
Moderate 300 0.5
None 49 0
Transformative 183 1
Unclear 4 0.125
Grand Total 783
Specificity how many weights
a. low 242 0.25
b. medium 292 0.5
c. high 394 1
None 30 0
Grand Total 958

Tags: ,

One thought on “An experiment with data from the open government partnership: Ranking countries

Leave a Reply