Florence house numbers import

From OpenStreetMap Wiki
Jump to: navigation, search

This page refers to the import of house numbers of the city of Florence in Italy. The data used is provided by the municipality of Florence.

Background

House numbering in Florence is peculiar, every street has red and black/blue numbers, the first ones are for commercial activities, the second ones for houses (at least at the beginning). For convention blue numbers are specified in a standard fashion, like 66, red number are specified putting behind an r or rosso like 66 r or 66 R. The two numerations are parallel and is possible that a road can have two equal numbers that differ only for the color. The numbers grow following the flow of the river Arno. This numeration system set some problems in map software that use interpolation of numbers, gps navigators can make mistake in the order of hundreds of meters in some situations.

The municipality of Florence start an interesting opendata project, among other datasets they provide a dataset with the totality of the house numbers accurately georeferenced and with the reference to the relative street. That dataset can be exploited to make OSM the most precise map for finding addresses in Florence.

Dataset

The dataset is avaiable at the opendata of florence portal, precisely at this page . The data is presented in different formats, the format used for the import is shapefile. That shapefile consist in a collection of punctual elements, one for each house number. each node has the keys:

  • ID an unique numeric identifier for the node;
  • CODICESTRA an unique identifier for the street of the house number, composed of numbers and letters;
  • CODICECIVI an unique identifier for the house number, composed from CODICESTRA followed from another code;
  • STRADA the name of the street of the house number;
  • NUMEROCIVI the house number (incluse the possible R and possible partitions);
  • TIPO the color, R for red, N for black/blue.

Import process

Basically 4 steps have been made for the import of the data:

  1. cleaning of the useless information of dataset;
  2. changing dataset key names in valid OSM key names;
  3. adding other keys,
  4. importing all the punctual elements of the dataset in points in OSM.

The first step has been made with QGIS, the others with JOSM.

In the first step the table of attributes of every point was cleaned deleting the columns:

  • ID;
  • CODICESTRA;
  • CODIC0ECIVI;
  • TIPO.

the key TIPO was useless because the color of the house number is included in the key NUMEROCIVI.

In the second step, for each point:

The values was keeped unchanged.

In the third step, for each point, was added the keys:

  • addr:city with value "Firenze" (Florence in italian);
  • addr:country with value "IT";
  • source with value "Comune di Firenze" (Florence municipality in italian).

In the fourth step was executed an upload of the points in bunches of 100 elements.

Result

The result of the import is good as you can see in the map, every street has is numbers.

Problems and improvements

There are two problems with this import

  1. some street names doesn't correspond in the house numbers and in the preexistents roads, i.e. "Via de' Medici" and "Via dei Medici";
  2. some preexistents address are duplicated with this import.

I'm planning to resolve the first problem with the help of OSM Inspector[1] adding an alt_name key to the streets not founded, I have maked a try with "Via de' Medici"/"Via dei Medici" and I'm waiting for the results.

Notes

  1. this is a direct link to OSM Inspector with correct settings for finding addresses with street not found in Florence.