Data Mining #data-science #question #content-management


Robert M. Taylor
 

I think this is the right answer and is the approach I've implemented successfully in the past.  Far more to do with process and content than any data analysis tool. I think that would be a far slower way to do it. 


Raquel Balceiro
 

Hi, Guillermo,

Good afternoon. Thank you very much for your explanations. For sure I’ll follow you up after taking a good look at the tools you suggested.

Regards and happy new year for you too,

Raquel Balceiro
Petrobras 

Em qui., 5 de jan. de 2023 às 05:29, Guillermo A. Galdamez <ggaldamez@...> escreveu:

Hi Raquel,

Happy new year!

When I read your message, I actually thought of a knowledge base or content management system as opposed to a data mining tool.

When you mentioned:
"Periodically, companies demand that we respond to these questionnaires, which are very similar to each other and whose answers are practically the same, only varying the data that we must provide depending on the accumulated history."

The first thing that came to mind was a Component Content Management System (CCMS). Heretto, which I will talk about a bit later, gives a pretty good explanation of what a CCMS is. In summary, the CCMS would enable you to assemble responses to questionnaires with reusable 'components' or chunks of information that are individually managed and stored in the CCMS.

If this sounds like it would be useful, I would recommend looking into Heretto and Kontent.ai.

Setting up CCMS solutions can be a bit complex, and would require you to spend effort into modelling your content (defining each type of component, and how you would describe it with metadata and taxonomies).  A simpler alternative may be something like Shelf, because even though it is targeted to customer service and call center companies, it will have some features that will cover your use cases (content reuse, albeit with less control/granularity as a full CCMS; and decision trees may be useful when covering for an expert's absence).

One final thing to note: You are right that you will most likely need a taxonomy - not only to keep a consistent count of the types of questions that you receive, but also to make your answers findable and reusable afterwards.

Full disclosure: Enterprise Knowledge, where I work, is a partner of each of these companies.

I hope you find this helpful! Good luck with your challenge. Feel free to follow up with any additional questions. 

Best,

Guillermo



On Tue, Dec 20, 2022 at 9:28 PM Raquel Balceiro <raqbalceiro@...> wrote:

Hi, everyone,

I can imagine everyone here is dealing with Christmas preparation, but I’d like to ask you for some help. I’ve been working with Compliance for seven years and, recently, we have been experiencing an issue with Due Diligence questionnaires. Periodically, companies demand that we respond to these questionnaires, which are very similar to each other and whose answers are practically the same, only varying the data that we must provide depending on the accumulated history.

We understand that using a Data Mining tool could help us easily answer these questionnaires, a task that today is very dependent on a single specialist. The idea is not to replace the specialist with the tool, but to allow other team members to easily perform the task when he is not available.

I'm not an IT person, so I really don't know where to start. I thought of creating a taxonomy with those questions, listing the frequency with which they are asked and the corresponding answer. At the same time, I thought of listing the sources of knowledge for each of these questions, as if I were building a Knowledge Map (people, documents, or systems).

Do you have any suggestions for a tool we could use? Something I should know about or do in this case?

Thanks in advance for your help.

Warm regards

Raquel Balceiro
Petrobras

--
Abraço,

Raquel Balceiro
(21) 99249-9287


Guillermo A. Galdamez
 

Hi Raquel,

Happy new year!

When I read your message, I actually thought of a knowledge base or content management system as opposed to a data mining tool.

When you mentioned:
"Periodically, companies demand that we respond to these questionnaires, which are very similar to each other and whose answers are practically the same, only varying the data that we must provide depending on the accumulated history."

The first thing that came to mind was a Component Content Management System (CCMS). Heretto, which I will talk about a bit later, gives a pretty good explanation of what a CCMS is. In summary, the CCMS would enable you to assemble responses to questionnaires with reusable 'components' or chunks of information that are individually managed and stored in the CCMS.

If this sounds like it would be useful, I would recommend looking into Heretto and Kontent.ai.

Setting up CCMS solutions can be a bit complex, and would require you to spend effort into modelling your content (defining each type of component, and how you would describe it with metadata and taxonomies).  A simpler alternative may be something like Shelf, because even though it is targeted to customer service and call center companies, it will have some features that will cover your use cases (content reuse, albeit with less control/granularity as a full CCMS; and decision trees may be useful when covering for an expert's absence).

One final thing to note: You are right that you will most likely need a taxonomy - not only to keep a consistent count of the types of questions that you receive, but also to make your answers findable and reusable afterwards.

Full disclosure: Enterprise Knowledge, where I work, is a partner of each of these companies.

I hope you find this helpful! Good luck with your challenge. Feel free to follow up with any additional questions. 

Best,

Guillermo



On Tue, Dec 20, 2022 at 9:28 PM Raquel Balceiro <raqbalceiro@...> wrote:

Hi, everyone,

I can imagine everyone here is dealing with Christmas preparation, but I’d like to ask you for some help. I’ve been working with Compliance for seven years and, recently, we have been experiencing an issue with Due Diligence questionnaires. Periodically, companies demand that we respond to these questionnaires, which are very similar to each other and whose answers are practically the same, only varying the data that we must provide depending on the accumulated history.

We understand that using a Data Mining tool could help us easily answer these questionnaires, a task that today is very dependent on a single specialist. The idea is not to replace the specialist with the tool, but to allow other team members to easily perform the task when he is not available.

I'm not an IT person, so I really don't know where to start. I thought of creating a taxonomy with those questions, listing the frequency with which they are asked and the corresponding answer. At the same time, I thought of listing the sources of knowledge for each of these questions, as if I were building a Knowledge Map (people, documents, or systems).

Do you have any suggestions for a tool we could use? Something I should know about or do in this case?

Thanks in advance for your help.

Warm regards

Raquel Balceiro
Petrobras


 

Hi Raquel, 

I use the open-source Orange Data mining (Link) where you can load 'add-ons' developed by universities or any 3rd party. 

In the context of the survey, I can think of these 3 add-ons in addition to data mining: 
- Sentiment Analysis is helpful when trying to construct an opinion or an affective state on a given subject. Perfect use for the voice of customers. 
- Text Analysis is helpful when visualizing co-occurrence, co-concordance, clustering, categories & classification across the survey topics.   
- Network Analysis is helpful when visualizing relationships among the responders (influence, decision-making, mentoring, expertise...) 

Based on the collected data and evidence, I would focus on the improvement areas and I would also try to highlight gaps, vulnerabilities, weak ties, and missing opportunities. It all depends on what the survey is trying to achieve and how it's designed. 

Thank you

Rachad 


Raquel Balceiro
 

Hi, everyone,

I can imagine everyone here is dealing with Christmas preparation, but I’d like to ask you for some help. I’ve been working with Compliance for seven years and, recently, we have been experiencing an issue with Due Diligence questionnaires. Periodically, companies demand that we respond to these questionnaires, which are very similar to each other and whose answers are practically the same, only varying the data that we must provide depending on the accumulated history.

We understand that using a Data Mining tool could help us easily answer these questionnaires, a task that today is very dependent on a single specialist. The idea is not to replace the specialist with the tool, but to allow other team members to easily perform the task when he is not available.

I'm not an IT person, so I really don't know where to start. I thought of creating a taxonomy with those questions, listing the frequency with which they are asked and the corresponding answer. At the same time, I thought of listing the sources of knowledge for each of these questions, as if I were building a Knowledge Map (people, documents, or systems).

Do you have any suggestions for a tool we could use? Something I should know about or do in this case?

Thanks in advance for your help.

Warm regards

Raquel Balceiro
Petrobras