User:Thomas.hills

Standardising & Entering Distributed Water Points

Motivation

Water & sanitation points are often recorded in OSM for a variety of reasons. This database is likely to be very useful in relief situations, but without a systematic tagging scheme it is difficult to quickly find the information required. Development and presentation of a suitable system could allow clearer tagging and more efficient understanding.

Summary

Thomas Hills has been looking at current tagging methods and possible improvements. Tallguy (Nick Allen) has been looking at suitable methods for cross-referencing water point data from external repositories such as the Humanitarian Water Point Data Exchange. This page is meant as a means of communication between OSM users and water point experts to develop a standardised water point tagging method, and a method of cross-referencing water point data using that method.

There is already a page on Water, Sanitation and Hygeine (WASH) on the OSM wiki but it hasn't been updated in a long time. The embryonic ideas there have been partially included in this page. There are HOT tagging presets for various water point types but these have not been updated for about four years.

Suggested project steps

As this is meant to be a collaborative process, this project plan is only a suggestion, and comments and criticism are welcome!

1. Observe & review current tagging methods to identify what water point information OSM contributors wish to enter into the map
2. Develop a preliminary standardised tagging procedure which accommodates the relevant information
3. Send procedure out for review by OSM contributors and water experts
4. Modify procedure based on feedback
5. Send preliminary water point cross-referencing procedure for review
6. Modify cross-reference procedure based on feedback
7. Launch water point cross-reference programme.

Current tagging methods

There is a great variety of tagging methods for water points, partly due to their heterogeneity and partly due to the lack of an existing clear tagging method. A survey of features with the tag amenity=drinking_water was performed across southern Africa - the results and discussion are shown on a [separate page]. Over 3441 water points, 156 different keys were used. Many were similar to those used on other nodes in the survey, and the tags could be grouped according to the kind of information they conveyed, as shown [here].

Preliminary proposal for a tagging protocol for water points in OSM

The table below sets out the current preliminary proposal for tagging water points in OSM. It is suggested that water points, used for more than simply drinking water, are tagged as amenity=water_point.

Key	Description	Values
OSM Location Keys
geometry	This specifies the type of feature	node way relation
id	The OSM ID
latitude
longitude
coordinates	This provides the lat/long of each point on a way.
addr:street	Used if it is not obvious where the access to the water point is from.
Specifying water point type
amenity	Identifies this feature as one which can provide drinking water	water_point
man_made	A range of water point types. Mutually exclusive to natural=*	water_well storage_tank water_tower water_tap sand_dam?
content	Used in conjunction with man_made=storage_tank	water
natural	A range of water point types. Mutually exclusive to man_made=*	spring water
water	Used in conjunction with natural=water	pond lake reservoir
pump	Explains whether there is a pump or not. More information on the pump can be given by specifying whether it is powered or manual	yes no manual powered
pump:type	This describes the model of the pump. There are several different manual ones, so this could be quite useful.
pump:fuel	I’m not sure about this, but I’d like to find a way to specify what fuel a pump uses, if it’s not manual.	diesel mains solar wind
depth	The depth of a borehole or well, in metres (Note, this would require a redefining of this tag - it is currently only used offshore)	[Number]
Ownership, access & fees
name	The regular name of the pump	[Text String]
ref	A reference name of the pump (perhaps some form of code, see Tanzanian rural waterpoint database for an example)	[Text String]
access	Used to describe the legal access	yes no
fee	Used for specifying whether a fee is usually charged for a service, or for access	yes no
charge	Used for specifying how much is charged for use/access to a facility	[Text string]
operator	Used to name an entity who is directly in charge of the current object	[Text string]
operator:type	Used to specify the kind of entity the operator is	private government ngo community
opening_hours	Specifies when the water point is available. There are rules regarding format
start_date	This is when the pump was constructed	Year
Availability
intermittent	Used to describe whether the water point is always open	yes no [Text string]
operational_status	Used to describe the current state of the water point	open closed
pump:working	Explains whether the pump is of operable quality	yes no
indoor	Is the pump inside or outside? May be useful in built-up areas.	yes no
sufficient_quantity	Is there enough water to satisfy demand?	yes no
Quality
condition	The condition of the water point, rather than the quality of the water contained therein	good fair deficient
water_purification	Whether water purification is applied/available	yes no chlorine
drinking_water	Is the water of drinking water quality?	yes conditional no
[Covered]	Specifies whether the water point is open to the elements or covered. However, covered=yes shouldn’t be used for underground elements.	yes no
Project & contact details
project	This is the name of the project.	[Text String]
contact:phone		[Text String]
contact:email		[Text String]
website		[Text String]
source	This is where the information came from. If it’s from a project on the ground, then it should be survey, not the name of the project.	survey
source:date
survey:date
note	Extra information	[Text String]

Questions to answer

• Should completely private water points be put on the map as an amenity? If they are private, they are not a community service/amenity.
• Should more address keys be included?

Review of Water Point Tagging in southern Africa

Summary

All 3441 points with amenity=drinking_water in Africa below 1°S were surveyed. 165 different keys were identified:

• 3 were present for 100% of features
• 6 were present for 20–80% of features
• 18 were present for 5–20% of features
• 37 were present for 1–5% of features
• 101 were present for <1% of features

The values attributed to these keys were studied and they were categorised into 52 groups according to function.

This is a preliminary survey of water points to try and develop a method to survey all the tags on water & sanitation points across the developing world.

Aim

The aim of this survey is to understand the requirements of HOT/OSM contributors when they add WASH facilities. This can then be used to develop a clear and useful tagging scheme which will cover most information which they may wish to enter.

Scope

This is only a preliminary survey and so covers only one tag (amenity=drinking_water) across a smaller region of the world. It covers sub-Saharan Africa from the Tanzanian Northern border (1.0°S) to the tip of South Africa, including Madagascar.

Method

A quick survey of the OSM wiki to identify common water point tags was performed. Those chosen for research were:

• amenity=drinking_water
• drinking_water=*
• amenity=water_source
• man_made=water_well
• man_made=water_tower
• man_made=storage_tank AND content=water
• man_made=water_tap

Due to there being so many water points in the region chosen, only features with the tag amenity=drinking_water were properly assessed. There were 3441 of these, and between them they had 156 different keys. An Overpass Turbo map of their distribution is shown on this page.

An Overpass Turbo map showing the distribution of amenity=drinking_water in southern Africa

The features were extracted as a geoJSON, converted to an Excel file using an online conversion tool ready for analysis. The results are shown below in tabular form, followed by a discussion.

Main Results

Table of Key Frequencies

This table shows the number of times each key was used across the 3441 water points.

Relative Frequency	Key	Frequency
100%	geometry	3441
	@id	3441
	amenity	3441
20–100%	latitude	3361
	longitude	3361
	name	1066
	watsan:type	909
	operational_status	874
5–20%	watsan:availability	622
	operator	617
	operator:type	525
	watsan:cost	509
	access	492
	water_point	474
	opening_hours	473
	watsan:water_public	450
	source	409
	man_made	397
	cost	363
	description	310
	source:date	236
	watsan:water_private	236
	watsan:location	225
	watsan:description	214
	watsan:users_per_day	207
	drinking_water	207
1-5%	pump	160
	watsan:water_availability	134
	construction_date	123
	owner	123
	ownership	123
	water_place_access	95
	status	92
	operator_type	87
	water_place_durability	87
	water_supply	86
	water_purification	81
	coordinates	80
	building	71
	building:levels	71
	addr:street	69
	building:material	69
	watsan:connection	66
	Operational_status	65
	note	64
	url	60
	watsan:connection_description	60
	ref	58
	operator:description	57
	project:eurosha_2012	57
	operational	52
	water_point_type	52
	water_quality	52
	water_quantity	52
	natural	49
	addr:hamlet	48
	addr:city	47
	addr:subdistrict	47
	start_date	42
	NOTE	40
	name:en	40
	watsan:location_description	38
	category	36
<1%	contact:phone	34
	media:video_device_number	31
	project	31
	source_date	31
	source_type_GPS	31
	hh_served	29
	charges	26
	damage_status	21
	media:camera_number	16
	functioning	15
	disused:amenity	15
	pump:status	15
	watsan:kwaho_pet_bottles	13
	watsan:kwaho_sodis	13
	watsan:kwaho_description	12
	watsan:kwaho_hwts	12
	place:district	12
	place:division	12
	place:location	12
	watsan:type_description	11
	operator:type_description	11
	watsan:toilet_public	11
	place:village	11
	media:camera_device_number	10
	watsan:funding	10
	pump:active	10
	media:video_number	9
	watsan:kwaho_health_products	9
	governance:facility_type	9
	governance:project_quality	9
	governance:project_status	9
	place:sub_location	9
	watsan:toilet_private	7
	watsan:storage_tank	7
	cosmha	7
	generator:source	7
	content	6
	opening_hours:description	6
	watsan:cost_per_20l	6
	watsan:litres_per_day	6
	watsan:water_availability_description	6
	Water_Point:operational_status	6
	governance:fund_source	6
	refurbished	6
	capacity	6
	landuse	6
	watsan:other	5
	fee	5
	water:availability_status	5
	indoor	5
	governance:operator	5
	contact:email	4
	watsan:bathroom	4
	contact:mobile	4
	water:water_point	4
	highway	3
	smoothness	3
	surface	3
	width	3
	watsan:cleaner	3
	place:sublocation	3
	Water_point:cost	3
	location	3
	governance:agricultural_borehole	3
watsan:water	3
notes	2
disused	2
image	2
watsan:other_description	2
watsan:features_per_location	2
description:de	2
governance:type	2
governance	2
governance:data_collection	2
designation	2
tank	2
website	1
operational_status:electricity	1
watsan:kwaho_health_products_description	1
governance:fund_quality	1
governance:fund_status	1
water:water_tank	1
source_type_map	1
wetap:status	1
wetap:statusnote	1
users	1
watsan_toilet:type	1
watsan:operator	1
	operational_status:status	1
	wasan_toilet:owner	1
	loc_name	1
	public	1
	private	1
	condition	1
	contact:postal	1
	governance:cost	1
	addr:neighbourhood	1
	watsan:water_pubic	1
	construction	1
	abandoned	1
	watsan:dumping_site	1
	name:fr	1
	name:af	1

Table of tag categories

This table shows keys which seem to convey very similar information, and are therefore suggested as being thought of as one category of information when developing a tagging procedure.

'Correct' key	Alternatives in the survey	Total frequency
geometry		3441
@id		3441
latitude		3361
longitude		3361
coordinates		80
amenity (alternative tags)	water_point designation category tank water_point_type water:water_point water_water_tank governance:facility_type governance:agricultural_borehole governance:type watsan:type watsan:description watsan:water_availability watsan:connection watsan:connection_description watsan:kwaho_description watsan:storage_tank watsan:water_availability_description	1989
access	private public water_place_access water_availability_status watsan:water_public watsan:water_private watsan:water watsan:other_description watsan:water_pubic	1288
operational_status	status Operational_status operational pump:status functioning pump:active damage_status wetap:status wetap:statusnote abandoned operational_status:status Water_Point:operational_status governance_project_status governance:fund_quality governance:fund_status	2848
name	loc_name	1067
operator	owner ownership operator:description governance:operator watsan:operator governance	926
fee	cost charges Water_point:cost governance:cost watsan:cost watsan:cost_per_20l	913
[Intermittent]	watsan:availability	622
operator:type	operator_type	612
opening_hours	opening_hours:description	479
note	NOTE notes description description:de watsan:location_description watsan:type_description	467
source	source_type_GPS governance:data_collection	442
man_made		397
source:date	source_date	442
[Users/Capacity]	hh_served capacity users watsan:users_per_day watsan:litres_per_day	249
drinking_water		207
start_date	construction_date	165
pump		160
project	project:eurosha_2012 cosmha governance:fund_source watsan:funding	104
[Purification]	water_purification	81
condition	water_place_durability	88
website	url	61
[Water & Point Quality]	water_quality governance:project_quality	61
ref		58
[Water Quantity]	water_quantity	52
natural		49
name:lang		42
contact:phone	contact:mobile	38
disused:amenity	disused	17
generator:source		7
content [of a storage tank]		6
indoor		5
contact:email		4
image		2
[Under Construction]	construction	1
operational_status:electricity		1
building		71
building:levels		71
building:material		69
addr:street	contact:postal	70
addr:hamlet	place:district place:division place:location place:village place:sub_location place:sublocation watsan:location	332*
addr:city		47
addr:subdistrict		47

Miscellaneous tags

watsan:kwaho_pet_bottles

watsan:kwaho_sodis

watsan:kwaho_hwts

watsan:kwaho_health_products

watsan:other

watsan:cleaner

watsan:features_per_location

watsan:kwaho_health_products_description

watsan_toilet:type

wasan_toilet:owner

watsan:dumping_site

watsan:bathroom

watsan:toilet_private

watsan:toilet_public

Irrelevant tags (for WASH)

media:video_device_number

media:camera_number

media:camera_device_number

media:video_number

capacity

refurbished

landuse

highway

smoothness

surface

width

location

Discussion of major keys with many alternatives

amenity

All points in this survey had the tag amenity=drinking_water, but many also had very similar tags which provided the same information. 18 alternative keys, with 1989 entries, were identified. This suggests that a comprehensive but useful tag describing the water point is required. Many values of the tags described a water tap, a water tank or a water tower, suggesting that people wished to provide more information about the type of water source beyond the fact that it could be drunk.

Note that, once other water point tags are surveyed (man_made=water_well, water_point=yes, drinking_water=*, etc.) more thought may be required to ensure that the new tagging procedure is comprehensive and covers their needs too.

8 of the alternatives are from one project, but give similar information to one another. Another project has 3 very similar ones. A clear tagging strategy in a WASH methodology could help projects such as these to efficiently capture and manipulate their data.

access

This is another popular tag, with 1288 instances of tagging across 10 keys. An interesting point with the alternative tags here is that 5 of the 9 alternatives are from one particular project, but seem to give very similar information. This suggests that there is a need for an access tag which can take a range of values.

operational_status

This tag had 15 alternatives, with 1165 instances of them being tagged. This is a critical tag for WATSAN projects and a suitable range of values could bring clarity to water point data. However, there is often the need to provide more information so a secondary category such as operational_status:note may allow for this to be accommodated.

operator

Several of the six alternatives to this had an owner-style tag, rather than operator. I think it can be assumed for most water points that the operator, rather than the owner, is the important piece of information. There were 926 instances of these seven tags being used.

fee

913 instances of this type of information were recorded across 7 keys. Values were a mixture of numbers and text strings, including currencies as well as costs per unit of water bought. It may be necessary to include two keys here, one which specifies whether there is a fee or not, and another with a text string where, for example, ‘7 KSh per 20L’ can be entered.

note

A mixture of ‘note’, ‘NOTE’, ‘notes’, ‘description’ and other tags were used 467 times across the points surveyed. The key ‘description’ is meant to be used to convey information to end users whereas notes are mainly meant to be read by other mappers. The conveyance of information kept within the ‘description’-style tags should ideally be able to be contained within the tagging protocol that is developed, and this tag can be avoided in the protocol.

Users/Capacity

There are quite a few keys which convey information about the amount of water available, or the number of people who use the water point. This is important to include within the protocol, and it may be split in two to produce better systematic values. There were 249 instances across 5 keys.

project

There are 104 instances of information about the project to which the water point belongs, across 5 keys. This is also important information, especially for finding your water points on the map using overpass turbo or a similar method. However, it is probably not necessary to have a separate key for each project, simply a value. It may also be sensible to include a project:date key to specify when the project was performed.

Addressing

Several water points were given addr: keys, like a building. Information about what settlement the water points are in should be clear from the map, but addr:street may be useful in built-up areas where the route to access the point may not be clear.