Merge pull request #467 from GFDRR/interview/asha

Asha's interview

docs/interviews/README.md

@@ -4,7 +4,7 @@
 
 - [Danny](danny.md)
 - [Stella](stella.md)
-- Asha
+- [Asha](asha.md)
 
 ## Summary
 
@@ -14,6 +14,8 @@
 
 > It is a (…) site where you can find the sources of various datasets, for specific countries, and their kind of openness.
 
+> The OpenDRI Index is a website that helps you upload datasets related to disaster risks in your country.
+
 ### What is valuable to them
 
 - **Data** there is a good amount of datasets
@@ -34,6 +36,10 @@
 
 - **UX** some terms are not always clear (global data, worldwide or multi-region scale datasets?)
 
+- **Dataset** category does not imply the many datasets will share a common ground in term of datapoints (column naming, data schema)
+
+- **Add a dataset** is effortful and the more dataset you add, the more time-consumming the maintenance of the metadata becomes.
+
 ### What could be made to make the Index more valuable
 
 - **Data** seeing water companies and university data as well
@@ -43,3 +49,7 @@
 - **Data** getting to more qualitative (experience of contribution, experience of exploration, experience of data wrangling)
 
 - **UX** label could convey their impact, what it is about, their usefulness and reading keys — instead of matching database field names (eg: "Dataset is open" vs. "Dataset can be downloaded publicly, without registration", etc.)
+
+- **Metadata** can be refreshed from the sources themselves (via a feed or a data connector of some sort)
+
+- **Metadata** could display in which real events the datasets have been used for (it increases the sense of usefulness)

docs/interviews/asha.md

@@ -0,0 +1,103 @@
+---
+date: 2019-03-13
+participants:
+- Asha Subramanian
+- Thomas Parisot
+observers:
+- Erick Gardin
+duration: P50M
+---
+
+# Why do you use the index?
+
+In order to submit India related datasets.
+
+
+# What do you remember from your last visit of the index?
+
+I remember the Index being a very transactional website, a website with a clear purpose.
+
+I remember the objective statement to be clear.
+
+I remember there are options to include datasets for your country.
+
+I remember there are instructions and informations about the categories of datasets.
+
+
+# What is your understanding of the OpenDRI Index?
+
+The OpenDRI Index is a website that helps you upload datasets related to disaster risks in your country.
+
+
+# Have you searched data on the Index?
+
+Yes. That's how I understood the datasets were not stored on the website, but only their metadata.
+
+❓ I wondered how the metadata were kept in sync with the sources?
+
+Aside of that, I was not looking for country data, nor any specific dataset.
+
+🗣 I was interested in discovering _how_ the data were organised.
+
+---
+
+🗣 I understand the OpenDRI Index collects the data, reflects their availability and tells us how is a government or country equiped to face disaster risks.
+
+❓ What about if a source adds new datasets on their website, or if the dataset change (location, name, formats, its objectives, etc.) or if a dataset is removed? **Is it reflected on the OpenDRI Index?**<br>
+⁉️ If so, there are high chances OpenDRI Index data to become inconsistent or outdated.
+
+> I question the kind of effort in submitting datasets.
+
+😒 It means additional work for me, to maintain the consistency and up to date informations.
+This is something that I need to be clear about with my management, to have that time to maintain the datasets metadata.
+
+😒 It is a _push_ effort, I need a motivation to do this work. What is this motivation to put the data on the Index?
+
+😒 I need to convince the agencies to open their datasets, most of them are closed. What is my benefit to do so?
+
+😶 There is no clear answer about that.
+
+💡 **In an ideal way**, I see these two things:
+
+1. there is a mechanism that **pulls changes** from the source website and the OpenDRI Index reflects those changes.
+I imagine a routine that figures out sources has changed, with some form of system notification;
+2. to compare countries, there must be a way to **compare contents** of datasets otherwise we are comparing apples and bananas.
+Two same datasets from different countries can contain datapoints which make them uncomparable. They must share enough common ground to be useable.
+
+# What is your understanding of the scoring system?
+
+I think i understood how it was calculated, based on known and existing datasets.
+
+🤨 I don't think they take in account the **usability** of datasets (coverage, trustworthiness, usefulness in case of events).
+
+😒 In a way, there is nothing about the reliability of the data but the 10 open data criteria.
+
+❓ So how do I convince my people to open their datasets?
+
+
+# Is comparing with datasets of other countries useful to you?
+
+Yes, it is always interesting to see what the best countries do. And why is that so.
+
+But it's hard to compare things which are not alike.
+
+❓Apart from rating the openness, why is the Index not addressing:
+
+1. the quality of the datapoints, so as the shared open data is really useful
+1. the everyday management of disaster risks by each country for the various types of hazards
+
+For instance, have the datasets been used to respond to disasters in the past?
+That link is missing.
+
+---
+
+❓ So if I spend time on the Index, what will I get out of it?
+If I share the datasets from our open data portal, I get some green ticks, but "how are they used" is the useful point for me.
+
+
+# Conclusion: how was the interview time for you?
+
+I really like that you are trying to _build_ and _reflect_ at the same time. I see so many cases where people get stuck in the building phase, without caring about the usability.
+
+I think it is useful and worthy to find out.
+