Comments
Big-Data & the Value Proposition
You must login to participate in this chat. Please login.

Not only an insightful presentation on "Big Data", and its applications, but a useful overview of the entire data gathering -> analysis -> insights process. Thank you.

Prospector

Thanks Tom, wondeful presetantion. In your own experience what would be the best Visualization tool to use?

Prospector

Thanks for archiving the presentation.

Prospector

There is a polarized opinion in our society on data visualization.  Trying to find common ground is very difficult.

Prospector

I don't think we have 'big-data' (at least in comparison to most) but this presentation really has giving me insight into our processes and the problems we are encountering.

Prospector

Needs Boudn Data Distribution - Our society wants to empower the staff to analyze data.  We use user friendly reporting tools.  The staff expects the software to be so plug and play that assumptions must be made when creating reporting universes - therefore problems arise when one assumption is made for a certain point-of-activity group which make another group's ability to use the tool problematic

Prospector

Open Analytics - the problem I feel we have is that a single contextual model drives the project and the data is gathering and stored with that in mind - making it harder to build a contextual model for another user group

Prospector

Cleaning data is very problematic in our society.  We find issues too late as well - usually when we are trying to perform Knowledge Extraction and by then we are pressed for time.

Prospector

This is our challenge - mining the online forumns when they are used.  But is it safe to make analysis on something that isn't used that much.  The analysis will be skewed towards those who participate and it might not be the real picture of our total customer universe

Prospector

The issue I see with semi-structured data collection is that even though we have fields, tables and work-flows set up for this it is hard to get the staff and customers to utilize this.

Prospector

Process Intelligence (or lack thereof) is a very important aspect of most of our data failures in new projects.

Prospector

And what problems do you run into when the data is created held and maintained by multiple departments (as well as the hardware)

Prospector

Look forward to the next seminars in the series - thanks A2

Prospector

Great presentation, thank you Tom.

Prospector

as stated below, visualization tools are a termendous help in sharing information with stakeholders and business owners

Prospector

testing assumptions and identifying relevance to the objectives leads to a successful engagement

 

Prospector

well defined goals and objectives are vital to the success of the projects

 

Prospector

Well presented, thank you.

Prospector

Goals, as stated, must be very clear at the beginning of a project.

Prospector

The amount of email may or maynot be an direct indication of value until testing is performed; assumptions have to be tested about all the diffferent forms of information.

Prospector

Visualization is a great tool for communication insights; it may now be a business critical tool.

Prospector

The Context of data is important and relative to the level of worker (line, management, exec's) and as well as the intended use; as stated the assumptions should be well documented and be auditable for management's review in the future.

Prospector

The Data Taxonomy is insightful to show the actual transaction information relative to the non-transactional information. Could the transactional information be a guide to how to interpret the amount of noise in the remaining 97% of the non-transformation information?

 

Prospector

Good information-glad i tuned in

Thanks TOM for the nice presentation

Data Doctor

data analysis algorithm can play a vital role for it

Data Doctor

but the process of converting raw data to results are important

Data Doctor

dats are more important for analysis

Data Doctor

thanks for the new initiative

Data Doctor

People have been most resonsive and have taken the time to give an indepth reply.

Prospector

Yan can easyly create a line to follow on your own readerboard.

Prospector

The blogs have added a new spectrum to the conversation.

Prospector

There are some great posts for follow up at the sugested sights.

Prospector

I've learned much from the questions asked by the attendees. Many of the answers and sharing  of experiences have added a timely insight to the quest.

Prospector

Cannot wait for the next one either.

Prospector

They always tell me something I did not know before.

Prospector

I really find these informative.

Prospector

Next Tuesday should be as good.

Prospector

I agree: this WAS a good one.

Prospector

Interesting information.

Prospector

Hope next week will be as good.

Prospector

This was a good seminar.

Prospector

good day to all

Prospector

Glad to be of help, @rnimmaka. :-)

 

Blogger

I hope everybody can join us for the second class in this All Analytics Academy semester. Next week Mark Pitts, data scientist and healthcare analytics executive (and frequent blogger here), will be joining us for a lecture, "Building the Big-Data Analytics Architecture." Wed., Feb. 20, 2 p.m. ET. 

Blogger

@Beth. I agree with your point. I think I would have to talk to my company people explaning the importance of having mutliple people with multiples skill sets. Really thoughtful Idea !!

Prospector

Thanks everyone

Prospector

@all; moving on to my next task; hope you enjoy the rest of the sessions!

 

Blogger

So it's time to wrap up here and say our goodbyes. Thanks Tom for a great lecture and to everybody for your questions. I know I've learned a bunch!

Blogger

Thank you Tom for a very informative class and chat, I have to be going now......Everyone have a great day !  : )

Blogger

@Tom, OK, I see that. 

Blogger

@eth that makes sense, as they would have insight into different angles of the questions addressed.

Blogger

My pleasure; thank you everyone for participating!~

Blogger

@Tom. Thank you very much for your repsonses. Really appreciate it.

Prospector

@rninmaka, to add to Tom's suggestion, I've heard of companies having really good results when they've formed teams comprising analysts, IT and business users. 

Blogger

@ariella: Better than I know humans!

Blogger

@beth.  Because people naturally try to contextualize to understand we communicate with each other without explicit context knowing the other can supply it.

@rnimmaka:  I think that every busienss unit needs to have an IT liaison person.  That person should be the "interpreter" who performs the interfacing.  That seems to work where companies have tried it.

@beth.  It does, but my point goes even further; think back to the fact that there are more favorable comments within six weeks of announcement.  The information there is in a separate bin, so getting the product release context and the social-network comment context to align is way more than formatting.

Blogger

Hey everybody, we're quickly approaching the top of hour and the end of this All Analytics Academy session. I think Tom is working on answers to a couple of the last questions. Did we miss any from anybody? Re-post if so, and if you still haven't posted a burning question, now's your chance!

Blogger

That, I imagine, goes to your point that if the data doesn't provide the structure your analytics has to?

Blogger

@Tom ah yes, you know Vulcans very well, as I recall.

Blogger

@Tom. Another question, how important is to work closely with IT or aligning ourselves to IT Enterprise Strategy especially when you want to leverage the advantage of big data technologies. I work in a marketing dept, where most of my senior staff are not at all conformtable talking to IT, They feel It always needs lot of time & effort to deliver something simple.

Prospector

@beth: Think comma-delimited spread-sheet files.  You have unstructured data but you have a delimiter and you have a schema, so it works.

Blogger

@batye:  And if you find a case where that's not true, we need to launch an analytics company to fix it!

@ariella:  True!

@Beth:  Sorry, beth, othe way around.  Most M2M is semi-structured because machines are more logical than humans (so are Vulcans!).  It takes intelligence to get things in a confusing format.  I don't think standardization in M2M is as important as simply consistent delineation of fields and a schema for interpretation.  It's a simple programming task to structure that sort of data.

Blogger

Tom, a question regarding M2M and its role in big-data analytics. I believe you said, while M2M is less an issue than the human element, that its many formats make context difficult to determine. Do we need some sort of standardization here? Or is the varied nature of M2M due to its use in varied industries?

Blogger

@Tom and admiinstrative access. Without that, they can't always get in to solve the problem.

Blogger

good point Tom :)

Prospector

General point:  Most IT is logical.  If you stop and think, any IT problem can be solved by an IT pro with access to reference material.

Blogger

@jchang, @tom, Nice question, even I face this same situation in my Org.

Prospector

@beth:  I think it requires a formalized decision management process, but that's not always to say a system is needed

@jchang:  That IS a good question; most companies start by making "collection" a policy and then move to working to include context data at the time of collection to add value better.

@rnimmaka My view is that it's always a major risk to launch a project based on a technology you don't have internal skill on.  In latin "Quis custodiet ipsos custodes", meaning "Who will watch the guards themselves".  You can't if you don't have in-house skill at least at the high level.

Blogger

@Tom yes, you'd have to be clear and not merely rely on inference.

Blogger

@Tom, what is your thought taking the help from cloud based vendors datameer when you dont have inhouse expertise although I am little worried of the security aspects?

Prospector

Thank you Tom, I always enjoy your presentations and learn a lot

Master Analyst

Another question - how does one break down the walls that naturally spring up in most organizations when it comes to data?  Even within my own org, we have walls of data, jealously guarded by their owners.  yet real value would come from being able to combine the data and search for relationships across the disparate domains.  Not just a big data question, I know....

\

Prospector

Tom, I couldn't help but to think as you were talking about processes during your lecture that big-data requires a formalized decision management system. Your thoughts on that?

Blogger

@noreen;  The majority of big-data projects are run in house (83% in part and 72% in total).  If you look at the number of successful projects, in-house wins over 9:1

Blogger

@Tom   Thanks well said, and exzactly what I was implying.......

Blogger

@Ariella:  Hi!  Right too.  The thing that's important is to not sound like you're a commercial, spouting canned data that all management suspects.  This should be about what the company needs to do, with other references showing it can be done and suggesting a mechanism.  Don't make it about the company in the case study, make it about YOUR company.

Blogger

Do most organizations stay in-house?

 

@Lewis:  I think in-house people are always better if we presume equal background, because they understand the organization, the data, the market/industry, and the politics.

Blogger

@Tom good point, the more specific and relevant the example, the better.

Blogger

I've heard that presenting insight visually gets the old mind going -- in other words, helps analysts or business users realize what questions or the type of questions they need to be asking. Would you agree?

Blogger

@rnimmaka:  The way that works best is to find industry-specific case studies or at least industry-relevant visualizations.  I find that telling your boss that so-and-so had good results doesn't move the ball typically, but if you show the graph that shows the decision/analysis chain that moved the ball THEN say how it helped, it means a lot.

Blogger

@rnimmaka   That is your opportunity to educate them !  : )

Blogger

MapReduce/Hadoop are really more about work distribution than about data analysis per se.  The general view of big data users is that you have to distribute queries and not collect data, and that's the model Hadoop provides.  The challenge is that you can't do that if you can't frame a query, and you can't really query anything without structure.

Blogger

each time I learn something new

Prospector

Tom, good example of insight vs. data. Thanks!

Blogger

@Tom, how does one explain the value of Big data if your senior is not really sure what big data is and how it can add value to your organization?

 

Prospector

@Tom   To add to jchang's and Beth question, this insight that is provided might be much easier to provide if you are an in -house analyst as opposed to a consultant ?

Blogger

@Tom thanks for your responses

@Beth - great addition to my question!

Prospector

@Beth.  I think all of the big vendors do a pretty decent job.  My personal opinion is to start with vendors who have a very strong visualization position first.

Blogger

Besides MapReduce and Hadoop, would you recommend other platforms?

Prospector

Tom, you mention reading literature from big-data vendors as a good starting point. Any big-data vendors doing a better job of delivering a practical context than others? in other words, any vendors to start with (from a background perspective, at least)?

Blogger

@Beth.  An example of "insight" is that favorable mentions of a product appear at a higher rate on a social network in the first six weeks of release.  An example of data is that there are 27 favorable mentions of products in August.  You can see the difference.

Blogger

@mzadroga the health care industry is a good example of an industry-specific value proposition.  You can take a high-level point like correctness of diagnosis, and you can spin it as something that's an input to patient-care outcome management, to managing insurance risk, to optimizing bed usage...you get the picture.  The same is true in the financial industry; loan performance, return on time of workers...

Blogger

Tom, let me ask a follow on question to jchang's last one, too, on data literacy. During the lecture you made the point that we should be providing the business with insight, not data. Can you share an example of the difference in delivery? How does insight appear to the business recipient vs. data?

Blogger

@Tom I see....I can imagine that some companies get "too excited" and don't really place a good structure in place before heading off into Analysis,  I can understand this, but it is probably counterproductive in the end ....

Blogger

@jchang;  All analysts start with exaggeration of credentials!  Seriously, the best approach is to read some basic material from big-data vendors so you get a practical context, and then try out what you've learned.  I believe in a build-as-you-go process

 

Blogger

Very informative.

Prospector

Thanks to everyone involved!  

Great presentation today!

Prospector

Tom- thanks for your comments.... any examples of industry specific value propositions?

Prospector

Tom - how does one best go about improving data literacy in the organization?  Seems we are relying more and more on data yet few folks seem predisposed to actually wanting to understand it.

Prospector


Tom, jchang asked earlier: How does a "regular analyst" get started with big data analysis?  Are there courses/programs that can be taken?  How does one translate coursework into real life situations and analysis?  What are the best ways to develop a toolkit for folks wanting to get into this field?

Want to tackle those questions?

 

Blogger

@Tom always a pleasure to attend one of your lectures.

Blogger
Thanks Tom and Beth. Super excited about the rest of the series - it is almost midnight in Dubai but will definitely be logging on every Wednesday the next few weeks,
Prospector

Thanks Tom, lots of good information !

Blogger

@Lewis Watson Some companies find it takes them months to establish a framework for contextualizing data, but once it's done it's not too hard to sustain

 

Blogger

I like Tom's point volume is not the major factor, it is corelations that can be derived that is most important. 

Blogger

What are some ways you would suggest to communicate to users the work required to provide some of their requests?

 

Prospector

Thanks for the great lecture Tom!

 

Blogger

OK gang I'm back on.  If you have any specific questions fire away!

 

Blogger

(sigh) I keep losing it.

Prospector

Thanks!  I think it was a temporary network issue on my end, seems to be fine now!

 

Prospector

No problem here either ....

Blogger

Does the audio keep dropping out for anyone else?

 

Prospector

@jrclark it's fine for me.

Blogger

I love the phrase, "otherwise it's just babble."

Blogger

Does the audio keep dropping out for anyone else?

 

Prospector

Applying structure where there is none, sounds like a lengthy process.....

Blogger

Hi jchang. Let's see what Tom has to say once he joins the text chat after he wraps up! Thanks for posting your question. 

Blogger

Hi Beth - I have a long question for Tom that I'd like his insights into.  My question is

How does a "regular analyst" get started with big data analysis?  Are there courses/programs that can be taken?  How does one translate coursework into real life situations and analysis?  What are the best ways to develop a toolkit for folks wanting to get into this field?

thank you.

Prospector

if you are having trouble hearing, please visit: More Troubleshooting

Prospector

 

If you don't see it or have trouble listening to it, we have a few troubleshooting suggestions:

- Try another browser

- Make sure you're running the latest version of Adobe Flash (get it here)

- Check with your network administrator; some corporate networks block streamed audio of the type this player uses

If none of these work, you may still be able to catch today's lecture via the archived version we will make available, and you can still participate in the text chat following the audio portion of today's class.

 

Blogger

I can't hear anything !

Prospector

Blocked by my firewall - will read through deck. Thanks...

Prospector

can't hear anything

 

Prospector

I am not hearing anything as well.

Prospector

I'm not hearing anything

 

Prospector

Yes, the program has begun.

Blogger

Hello every one

 

Prospector

Hi Beth!.. always nice to see a Familiar Name!  ;)

Prospector

Going off the board now until the chat after the preso!

Blogger

Beth and I go way back so I can be playful!

Blogger

Hi, Beth; I'm sure you know everything in the industry already!

Blogger

Hi Tom. Looking forward to the lecture!

Blogger

"Netweaver", huh.  We have a tech journal (in its 32nd year) called "Netwatcher".

Blogger

Good morning or afternoon all

Master Analyst

Hi, everyone, and thanks for attending!  We'll be starting shortly with the usual format of a presentation followed by a chat here.  Be sure to download the slides; I'll be telling you when to go to the next one.  I'll be off the board during the preso but back on for the Q&A.  Forgive me if I have a slight cough too!

Blogger

only 25 more minutes to wait!

Prospector

Students, a quick note before we begin. If you haven't already downloaded today's lecture, please do so by clicking on the "Today's Slide Deck" link above.

Also note: the audio player for this lecture will open at start time. If you don't see it or have trouble listening to it, we have a few troubleshooting suggestions:

- Try another browser

- Make sure you're running the latest version of Adobe Flash (get it here)

- Check with your network administrator; some corporate networks block streamed audio of the type this player uses

If none of these work, you may still be able to catch today's lecture via the archived version we will make available, and you can still participate in the text chat following the audio portion of today's class.

 

 

Blogger

when are they going to update the live chat display?

Statistician

Greetings earthlings

 

Prospector

Hope this one is at least as good as the last one.

Prospector

Cannot wait for this to begin!

Prospector

Tuned in and ready!

Prospector

Looking forward to it!

Statistician

Hello from Sunny San Diego...

Looking forward to the discussion.

Thanks

Prospector

Hellow All, Hope to attend

Data Doctor

hope to attend

Prospector

Great to be here!

Prospector

I appreciate your expeditious.

Prospector

see you all next week!

Prospector

Looking forward to this discussion.

Prospector

Looking forward to another learning experience.

Prospector

Looking forward to  February 13

Prospector

Looking forward to learn about the value proposition

Prospector

I look forward to this discussion 

Prospector

Don't want to miss this one.

Prospector

We're absolutely thrilled to be launching our first series of online classes, and so happy to be welcoming Tom Nolle to All Analytics as an instructor. Tom and I go way back, and one thing is for certain: He knows his stuff! 

Blogger

See you all on the 13th... have a Great Weekend!

Prospector
Looking forward to it. Cy'all on the 13th.
Prospector

It will be a cool session!

Prospector
I'm excited to see Tom Nolle is the first lecturer! He's a great presenter who always delivers exceptional inforamtion. I hope to see a lot of participation in these All Analytics courses.
Data Doctor


INFORMATION RESOURCES
ANALYTICS IN ACTION
CARTERTOONS
VIEW ALL +
QUICK POLL
VIEW ALL +