File:Opasnet base connection.ANA

This model is an interface for uploading data to and downloading it from the Opasnet base. This file is an editable version of a web interface. For using the file, you need Analytica Enterprise. For using the web interface, you only need a web browser (Mozilla Firefox recommended).

An improved web interface

There is also a web interface for downloading only; this interface produces raw data tables that you can copy-paste to an external software. However, some pieces of data are very large, and you may experience downloading the data, or handling the data afterwards. Therefore, there is a need to improve the interface. The improvement needs are described in this chapter.

The data is stored into the Opasnet base as two-dimensional tables, one for each data object (data objects can be either studies or variables. For details, see the description of the database structure. For example, the data called Morbidity data for Europe consists of a table that is indexed by four indices: Year, Sex, Age, and Country. Indices define the specifics of a particular observation. In addition, there is one parameter (morbidity), which is the field that contains the observations (values of interest).

Often, the user is only interested in a part of the data, e.g. one year of observation, one country, and maybe a few age groups. So, there should be an interface that shows the data in the Opasnet in a few steps, when a user selects a data object.

1. The user is shown the list of uploads (versions) of the data object. The default is the newest upload, but the user should be able to choose an older one as well. This is based on the query that is shown on the metadata box as "List of uploads".
2. The user is shown the list of indices and locations (possible values). The user can choose any combination of locations with tick boxes beside each location.
3. The user is shown a two-dimensional table of data with indices and parameters as fields, and observations as rows. This can then be copy-pasted to other programs. Alternatively, a comma separated file (CSV) is created and a Save as dialog box appears. Now the "The newest upload" link does something like this but does not allow for selecting locations, and the output format is not optimal for copy-pasting data into another software.
4. In cases where the data is given as distributions instead of point estimates, the data can also be downloaded as a sample file of distributions.

Development needs

• Now the model interprets the results as numbers if applicable, otherwise text.
• Analytica_node creates an indexname list where local indices start with dot (.Index). The dot should be removed, because in the Opasnet Base there are no local indices.
• Upload the read functions. The indices should have the real names from the Opasnet Base, not just In1, In2 and so on.
• The queries for read functions should be the same queries as in [[Template:Resultlink]].
• The read functions don't work properly. They don't show Testvariable data uploaded 28.8.2009 for some reason.
• The first read function should show data in 2D format; only the next function should turn it into an array. This is because not all data has originally been arrays and would result in hugely multidimensional arrays with a lot of empty cells.

Kehitettävää (by Erkki):

• A) Taulukkodatan syöttö on vielä epävarmaa, toimii välillä ja välillä ei, vailla selkeää johdonmukaisuutta. ==> Epävarma toiminta voisi Juhan mukaan viitata johonkin selain-/flash-/tms.-tekniseen fibaan.
• B) Parametrisarakkeissa olevan puuttuvan datan käsittely näyttää keskeneräiseltä, koska 'Null' ei nähtävästi kelpaa arvoksi lainkaan (ko. rivit jäävät kokonaan pois), ja 'Undefined'-arvot AWP muuntaa nolliksi, mikä johtaa harhaan.
• C) Kun menen katsomaan http://base.opasnet.org/, niin ko. study näkyy kylläkin pudotusvalikossa, mutta "post"-painikkeella ei palaudu mitään dataa. Juha varmaan fiksailee tuon kun ehtii.

See also

• Previous version image:RDB connection.ANA and image:RDB reader.ANA were only for either reading or writing data.