This report on tourist accommodation and touristification in Universitat de València is a joint initiative of Montera34 and the Institutional Chair of Collaborative Economy of the Universitat de València and is the result of the workshop held in March 2019.
Given the lack of data on tourist housing we offer open data, open analysis methodologies and visualisations. This workshop and report are part of a series of activities organised by Montera34 called Airbnb Effect.
This report is the result of the Participatory Science Workshop on the Airbnb Effect in Valencia from 1 to 3 March 2019 at Las Naves in Valencia.
The workshop was coordinated by Montera34, using a methodology that allows learning how to obtain, prepare and visualise data while producing a rigorous collective analysis on a real and local case.
This project arises from the lack of transparency of Airbnb data, the most popular tourist accommodation rental platform. Faced with this opacity, we offer to open up the data we have managed to obtain, document methodologies for open analysis and create visualisations to understand the current situation. We have also carried out this analysis collaboratively in the form of a workshop.
a. Working groups
After a morning of theoretical presentations during the one-and-a-half-day workshop, the following working groups were formed to try to answer the following questions:
Which, in what proportion and where are the officially registered dwellings?
Counting words: BEACH and CENTRE in the text of advertisements.
Areas of vulnerability and tourist flats
Do tourist flats contribute to the rehabilitation of dwellings that were in poor condition?
Other issues analysed
The report has been completed with other sections to complement the information:
Concentration of accommodation in users
Seasonality
Evolution of Airbnb's presence
b. Participants
This report is possible thanks to the work of the people participating in the Airbnb Effect workshops.
On 27 February 2019, there were 6,552 listings published on Airbnb in Valencia, with the capacity to accommodate 25,537 people.
These listings (listings in Airbnb terminology) have been published by 3,842 different users (hosts in Airbnb terminology).
2,236 listings specified a registration number as a Holiday Home, which is 49% of the total number of complete properties advertised. This is not to say that all of these registration numbers are valid.
How ownership and management of the advertisements is concentrated
As explained in the methodology, we differentiate between host users and users with two or more listings:
Users who manage an advertisement. These could be included in a collaborative economy model in which they rent out their usual home, or part of it. They can be considered to be making an asset available when they are not using it.
Users who manage two or more ads. When more than one advertisement is managed, it can be considered that at least one of the accommodations advertised is not a regular dwelling, so the activity of these users cannot be considered a collaborative economy.
On the other hand, it should also be noted that it is not possible to differentiate between homeowners and managers: there are Airbnb users, usually companies or professional managers, who manage accommodation for multiple homeowners.
There are 2,858 users who manage a single listing (74.4% of all hosts), compared to the remaining 25.6% who manage two or more listings. Users who manage a single advert offer 43.6% of all accommodation, which is 43.8% of the total number of bedplaces.
Multi-managers (25.6%) control 56.4% of the accommodations and 56.2% of all places offered on Airbnb. Specifically, those who manage 3 or more listings (11.2%) control 10,502, 41.1% of all places.
13.0% OF THE TOTAL NUMBER OF HOSTS WITH MORE PLACES OFFER 46.8% OF THE PLACES (500 USERS MANAGE 11,941 PLACES).
Comparison with Madrid and Barcelona
Comparem les ciutat de València amb Barcelona i Madrid (amb data de setembre de 2018) i mateixa font de dades, InsideAirbnb. València ronda els 6.500 anuncis publicats enfront dels més o menys 20.000 de Madrid i Barcelona.
Valencia has 74.42% of hosts with only one ad, compared to Madrid 77.2% and Barcelona with 72.3.
3.8% of hosts with 5 or more ads have 27.9%. Barcelona 4.8 : 46.2 and Madrid 4.4 : 36.3.
That is, for more or less the same percentage of hosts with more ads (5%), Barcelona hosts manage 10% more ads, approaching 50%.
The distribution of the concentration of ad management is in any case similar:
about 75% of the hosts have a single advert managing 30-40% of the places.
multi-manager hosts account for 25% of the hosts and manage 60-70% of the places.
Top 10 hosts
These are the top 10 hosts, or maybe we should say better, the top 10 vacation home managers with the most listings on Airbnb. They probably have more in their portfolio but these were the ones collected by the scraping in September 2018. We show the number of accommodations and places offered by these top 10 managers in Barcelona.
On aqueix situen els allotjaments dels top 10 per nombre d'anuncis?
The top 10 hosts (0.1% of all hosts) with the most places have 5,201 places available (which is 8.2% of the total number of places).
The top 20 hosts (0.2% of all hosts) with the most places have 7,806 places available (which is 12.3% of all places).
The top 50 hosts (0.5% of all hosts) with the most places have 13,239 places available (which is 20.8% of the total number of places).
The top 100 hosts (1.0% of the total number of hosts) with the most places have 18,046 places available (which is 28.4% of the total number of places).
The top 200 hosts (2.0% of the total number of hosts) with the most places have 23,445 places available (which is 36.8% of the total number of places).
The top 300 hosts (3.0% of the total number of hosts) with the most places have 26,753 available places (which is 42.0% of the total number of places).
The top 500 hosts (5.0% of the total number of hosts) with the most bedplaces have 30,932 available bedplaces (which is 48.6% of the total number of bedplaces).
How many tourist homes are published with and without a registration number in the city of Valencia? How are they distributed by neighbourhood according to Airbnb data?
How many tourist dwellings are registered with the Department of Tourism compared to those published on Airbnb?
What is the concentration of these dwellings by neighbourhoods, identifying how many are registered and how many are not registered?
How many listings have a registration number?
At the end of February 2019, 6,552 ads were published on the Airbnb platform, of which 4,561 correspond to complete flats that are, in legal terms, Tourist Dwellings. We are going to study exclusively the complete dwellings for this analysis.
Of these 4,561 Tourist Dwellings, in a first analysis through the field "License" we detected that there were 2,104 with a registration number, 46.1%, and 2,457 without a registration number, 53.8%.
The data obtained caught our attention and we realised that not all hosts put the registration number in the corresponding section, but rather in different sections of the advertisement, for example in the title of the advertisement ("name") or in the description or summary ("notes", "summary", "description" or "space" fields). For this reason, a search for patterns coinciding with registration codes in the aforementioned fields was carried out, initially the pattern was set at 5 digits corresponding to the generic registration number.
After this process, the registered tourist dwellings amounted to 2,104, which corresponds to 46% of the total of the 4,561 dwellings; the unregistered dwellings amounted to 2,457 (54% of the total).
In a subsequent phase, in order to detect how many of these 4,561 dwellings were active, we filtered out those with a review number other than 0 and a host response rate ("host_reponse rate") also other than 0. Under these criteria, the number of tourist dwellings was 3,908, of which 2,003 appear with a registration code published in one of their sections, which represents 52% of the total number of active dwellings and 1,905 dwellings without registration (48%).
Summary
Airbnb listings
Type
Number
% of complete dwelling
Total advertisements
6.552
-
Complete housing advertisements
4.561
100%
Ads complete houses with licence
2.104
46,1%
Ads complete dwellings with revised licence
2.236
49,0%
Listed on Airbnb and with number of reviews higher than 0
Type
Number
% of complete dwelling
Total advertisements
5.340
-
Active complete housing advertisements
3.908
100%
Active ads complete dwellings with revised licence
2.003
51,2%
Distribution by neighbourhood
The graphs show the distribution of the above results by neighbourhood.
Official Register
Let us compare this data with the information from the Generalitat Valenciana, the Official Register of Tourist Dwellings in the Consellería de Turismo. According to it, there are 5,809 dwellings registered in Valencia.
As the study is incomplete, we cannot affirm that the advertisements published without a registration number mean that they do not have one, we would need to compare them with the data from the Generalitat to be able to draw conclusions, this fact only means that it does not appear in the advertisement.
Conclusions
At this point, it cannot be affirmed that the advertisements published without a registration number mean that they do not have one, nor that those that have it published are indeed registered dwellings. What we can say is that there are at least 32-34% of properties that are active and do not indicate their registration number.
Errors detected: until a year ago, Airbnb did not have a section for listing the licence number, so many properties do have a registration number but it is listed in a different section, such as in the description, title or other notable aspects, which makes the study difficult.
The host has, as one of its most effective marketing strategies, the possibility of including in the descriptions of its advertised properties certain messages that include the identifying concepts of the district or area. In this way, we would expect to find a specialisation and differentiation of the advertisements in each district. Thus, we would find that maritime districts include concepts such as "beach" or "sea" or that central districts tend to emphasise their status as the "heart" of the city or as a "central" or similar place.
This exercise analyses the words contained in each advertisement.
Future steps
In a second phase of this study, one could, for example, carry out a more detailed analysis of expressions rather than words. For example to distinguish "centre" from expressions such as "close to the centre" or "5 min walk to the centre". In this way, the maps would be more accurate. Also, in order to avoid that a district is automatically filled in, it would be more accurate to mark with a dot the flats containing the word "centre" and to draw polygons of influence.
How much presence of tourist flats is there in areas of vulnerability in the city of Valencia?
This exercise analyses two different indicators of urban vulnerability, from the City Council and the Valencian Regional Government, and compares them with the presence of Airbnb in neighbourhoods in Valencia. It is not intended to imply causality between the variables, but rather an analysis tool to demand or implement measures as a priority.
Vulnerability index and Airbnb presence in neighbourhoods
The ratio of available Airbnb places per 100 inhabitants is used to compare with the vulnerability indexes.
According to the "Sensitive Urban Areas" of the Valencia City Council
Since the vulnerable areas indicated in each of the indices vary substantially, the priority areas of study that coincide would be:
El Grau
Beteró
Morvedre
Central area of Botánico
La Roqueta
And those to be added:
El Carmen
Velluters
Malvarrosa
East area of Ruzafa
According to the vulnerability index of the Conselleria
An analysis of the overlap shows that there are some neighbourhoods where the impact of the concentration of accommodation advertised on Airbnb on the vulnerability of these areas should be studied as a matter of priority. Specifically in the following neighbourhoods:
El Cabanyal-Canyamelar
El Grau
Beteró
Aiora
West Benimaclet area
Morvedre
Central area of Botánico
La Roqueta
It would be important to study in particular those areas classified as "residential vulnerability" in order to demand or implement measures as a priority.
Number of inhabitants by census section (Source: INE www.INE.es) To obtain the table with the continuous census of inhabitants: www.INE.es > [(Left Bar) Demographics and Population / POPULATION PER MUNICIPALITY / Continuous census statistics: Detailed information / (Left Bar) Results / Tab: Definitive / Detailed results. Period 1996-2017 (By census sections: Year 2018) / (From the list we select the province) Population by sex, section and age (five-year groups) / (Selection as in lower view and) Consult Selection / (Top right) Download icon / Format: Excel extension XLS)] > 4601.xls. From the whole Valencian Community, only those corresponding to the city of Valencia will be selected (THOSE BEGINNING WITH 46250), which will allow all the information to be downloaded in a single Excel file, as less than 10,000 cells are contemplated.
Ideally, we would have been able to count the number of refurbished properties per neighbourhood, obtain a percentage of the total number of properties per neighbourhood and compare them. However, there were very few refurbished plots in the last ten years (a period chosen to coincide with the number of years that airbnb has been active): 300 refurbished, out of a total of 38,000 plots in the city. This leads us to believe that the data is not sufficiently up to date or that there is insufficient monitoring. In any case, this register only includes major refurbishments, which have required a building permit, and on which the cadastre has carried out a revision-revaluation once the work has been completed.
To access the cadastre data you have to access the electronic headquarters of the cadastre (https://www.sedecatastro.gob.es/ -> "DIFFUSION DE DATOS CATASTRALES") with an electronic certificate or PIN code.
Download .CAT file -> manual to transform the CAT file into csv that can then be joined in QGIS with PARCELA.shp using the field REFCAT (parcel) with 31_pc (type 14): http://www.catastro.minhap.es/ayuda/manual_descargas_cat.pdf In this way we obtain a PARCEL layer with a column with title 75_ar that indicates the year of reform (in case there is reform).
