Ask-JGI queries#

How they work#

  1. Researchers email their queries to ask-jgi@bristol.ac.uk.

  2. An automatic email is sent to the querier, as explained here. This email asks the querier for additional details (e.g. description of aim and problem, description or excerpt of data) and letting them know to expect a response within the next few days.

  3. The Ask-JGI student on shift will share the query in the Ask-JGI Teams channel, and will either:

    • confirm that they intend to reply to the query, or

    • ask whether anyone else is available to do so. Noting the new query in the Teams channel gives those not on shift the opportunity to collaborate on the query if it is of interest to them, or falls within their own area of expertise. Likewise, the person on shift may not feel that they have the appropriate level of expertise in the relevant area. If this is the case it is also possible to ask for advice from the rest of the team on how to proceed, or to request support to answer the query.

  4. Once it has been agreed who will pick up the query, that team member is responsible for assigning the task to themselves on the Ask-JGI tracker, and recording information on the and faculty of the querier, and allocating relevant tags. Instructions on how to fill out tickets on the Ask-JGI tracker can be found here.

  5. The Ask-JGI team member responsible for the query will respond to the querier to either answer their question or to arrange a meeting to discuss further. It is important to reply to the latest email in the email thread (which includes the Ask-JGI automatic response) so that the Ask-JGI ticket number is included in the continuing email thread. This ensures that the same tracker ticket is updated as the email thread evolves. For more information on this, see here.

  6. The Ask-JGI team member should use the notes section of their query ticket on the Ask-JGI tracker to record the status and current actions from their query.

  7. Once the query has been resolved, the ticket on the tracker should be marked as resolved.

Note

You can store any files related to your queries here. Remember to give any folders an informative name, so future shifters can access your files.

Query Categories#

The tracker contains fields for “Query topic” and “Task for Ask-JGI”. These are so that the JGI can track what kind of query we recieve and what kind of help we offer most often. The different options for these fields are described below - if you are unsure which one to use, you can ask the Teams channel or the team in the weekly meeting.

Query topic#

This refers to the broad area(s) of data science that the query is about. There may be more than one - e.g. a query may be about both Statistics and ML. The current options in the tracker are:

Option

Meaning

Experimental Design

Looking for ideas on what to do with data, or how to best investigate something using data

Ethics

The ethical considerations involved when dealing with data and data science techniques

GIS

Geographic Information System - usually related to maps

ML

Machine Learning

Natural Language Processing

NLP - Processing human language

Privacy/Security

Ideas about data security and privacy

Statistics

Classical Statistics - i.e. making inference using a data sample, but not ML

Support with application/bid

The JGI can offer letters of support and/or support applications and bids by University researchers

Talk/promotion

A talk/event/promotion that may be of interest to the JGI

Visualisation

Help with data visualisation; can be either coding support or advice

Web Scraping

Automatically retrieving information from the web

Task for Ask-JGI#

This refers to the type of action that will be undertaken by the Data Scientist(s) assigned to this query. There may be more than one - e.g. a query may require both Technical advice and Signposting. The current options in the tracker are:

Option

Meaning

Technical advice

Offered technical advice - on e.g. frameworks, languages, tooling, methodology

Signposting

Pointed to a relevant resource

Comments on document

Commented on a document sent by the user

Non technical advice

Offered non-technical advice - on e.g. sharing ideas, funding suggestions, networking

“Rubber ducking”

Helped debug by acting as a sounding board

Quick email response only

Query resolved after a quick email response (or a few short emails)

Writing code

Provided code for the user - either an example for them to learn from or a prototype