-
Notifications
You must be signed in to change notification settings - Fork 12
About this project
What is this amount really worth ?
That's the simple question this project wants to answer.
News and political speeches contain a lot of big amounts but such numbers are not easily understandable by themselves.
Can you really tell what's 5 billion US dollars?
We created this tool to compare such big amounts, we called these amounts the stories.
A story, in our application, represents an expenditure item. It can be all kind of expenditures, like public spending, construction costs etc.
Every spending must be described with the following information:
- the currency used
- the amount of the expenditure
- the year
We identified two different kinds of stories: unique events and repeatable events.
Unique (or discrete) events are spending that are set for a particular event such as the construction cost for a building. They're not meant to be repeated in the time, when the building is constructed the story ends here.
These unique events can be compared to an amount thanks to a percentage.
this amount is X percent of this story.
Repeatable events (or "over one year" events) are events repeatable over the years. The annual government's budgets are a good example of repeatable events.
When we compare an amount to such stories we can give a time equivalence (like months, weeks, days):
this amount is X months of this story.
Sticky stories are not a different kind of stories. It's just a way to make some stories more important in our application. Such stories are the stories the administrators intend to draw attention to.
This can be well known stories that are easily memorable.
Because we are dealing with past events and spending amounts, we need to compensate the inflation of the stories. Indeed, in order to compare big amounts we need to make them comparable. To ensure every stories can be compared to an amount we also convert their amount into US Dollars.
We use 2 information to do this compensation:
- the story's consumer price index (inflation level) based on its country and its year.
- the conversion rate of the story's currency to US dollar.
- Consumer price indexes: World Bank, through a fork of the economics python module.
- Currencies exchange rates: OpenExchangeRates
We introduced a feature in our application to get a list of stories sorted by relevance (check Get results sorted by relevance to see how).
When we say that a story is relevant compared to an amount we mean that the story's amount compared to the user amount produce a nice equivalence.
- For unique stories it will be a nice percentages like 50%, 90%, 10% etc.
- For repeatable stories the relevance will be high if it's a round number of months, weeks or days.
This equivalence & relevance system is really useful to retrieve all significant stories compared to an amount. Therefore we are able to answer our main question: What is this amount really worth ?