dataGecko loves working with data, but even more, dataGecko likes to empower others to turn their data into information they can really use. dataGecko has over 20 years hands on experience working with data, creating systems to analyse it, and delivering results to end users. There are few data sources we can’t work with, but every situation is unique and requires it own approach. dataGecko’s strength lies in having significant experience working in various sectors within the information universe, especially corporate data analysis in the areas of IT and telecommunications, academia, and library and information management. Let’s start with an example of the type of work we do.
Data analysis of Survey Results.
How do I get more meaningful data from my survey results?
Online surveys are often posted with grand plans to answer all your questions about a particular subject, but more often then not, what follows is frustration, because the results provided are usually too generalised to be really useful, and the real detail is hidden from view. You can be literally swamped with data, but have too little useful information. This is because, with most online survey systems you get back results based on the overall summaries of the respondents, which is sort of like buying a cake but only getting the icing on the top. It is important data, but what you really would like to do is get at the goodness of the cake underneath. This is what the idea of “business intelligence” is all about, slicing up the cake in as many ways as you can imagine so you can see the detail insidefrom any possible perspective. “Slicing the cake” is the dataGecko’s speciality. (We quite like eating it as well!)
What does this mean actually?
Well, imagine you have collected data in your survey about the number of customers in each of your major operational locations. You might get a result that looks like this:
|
|
Total |
% |
|
Melbourne |
41 |
37.6% |
|
Sydney |
31 |
28.4% |
|
Brisbane |
15 |
13.8% |
|
Canberra |
15 |
13.8% |
|
Adelaide |
3 |
2.8% |
|
Hobart |
3 |
2.8% |
|
Darwin |
1 |
0.9% |
|
|
109 |
100.0% |
But what if you also asked if your customer’s wanted to attend an information session about your latest product or service, you might get this sort of response.
|
|
Total |
% |
|
Yes |
50 |
45.9% |
|
No |
59 |
54.1% |
|
|
109 |
100.0% |
Maybe you would think that there is not much interest, as less than half the respondents indicated this. But wait, is that the whole story? Maybe some locations have more interest than others? How do you know which cities the interested respondents are in? You need to break down the responses by city, but your survey software doesn’t often provide this sort of breakdown. Your only option may be to download the raw data and try to work it out from the individual responses.
What dataGecko can do for you is load your data into a simple to use dynamic data analysis program, in a report customised to your particular survey, and then you can use this to see what the actual breakdown is.
|
Attend info session? |
Total |
% |
Yes |
% |
No |
% |
|
Melbourne |
41 |
37.6% |
36 |
88% |
5 |
12% |
|
Sydney |
31 |
28.4% |
3 |
10% |
28 |
90% |
|
Brisbane |
15 |
13.8% |
7 |
47% |
8 |
53% |
|
Canberra |
15 |
13.8% |
1 |
7% |
14 |
93% |
|
Adelaide |
3 |
2.8% |
2 |
67% |
1 |
33% |
|
Hobart |
3 |
2.8% |
1 |
33% |
2 |
67% |
|
Darwin |
1 |
0.9% |
0 |
0% |
1 |
100% |
|
|
109 |
100.0% |
50 |
|
59 |
|
Right away you can see that Melbourne respondents are actually very interested in your information sessions, something the actual survey didn’t easily tell you. This is the difference between information and data. Of course this is a very simplistic example, and you are thinking “but I could do that in Excel”. You could, but for each interrelated set of data you would have to manually extract the information and recombine it before you could perform this analysis – tedious work. With our system, once the data is in the analysis program, you can do this type of “slice and dice” analysis by ANY question (or group of questions) and for ANY set of answers (or subsets of answers) within the survey – quickly, easily and intuitively, with no additional work! Its as simple as picking options from some drop-down boxes. (See below for some demonstration screen shots of this is action.) Because the data is now in a BI datacube, this type of comparative analysis becomes child’s play. Maybe you want to see all the survey results grouped by age, experience level, years of service, income, or for that matter, any other category-type question you have in the survey. dataGecko’s solution allows you to view the data by any number of these layers, separately or combined, and unlike most analysis packages, it takes only seconds to move from view to view. Assuming you have the right questions in your survey, you can quickly answer complex questions, such as “show me the interest in attending a session only for people in Brisbane or Sydney, who have more than 10 years of service and that have a budget of over $5,000”. No complex syntax is required, just select a few filters. Once these filters are set you can view all the responses in the survey based on this subset of responses for every other question in the survey! This is achieved with just a couple of intuitive mouse clicks, in seconds. It’s easy and anyone can do it. This is data analysis the way it was meant to be! You are truly free to explore your data without the mechanics of the system getting in the way. You really have to try it to believe how easy it is.
How is the data analysis achieved? What software do you use?
If you have ever had to perform some data analysis, you may have tried a number of statistical analysis programs, such as SPSS, or tired to do your own analysis in Excel. Either way, its hard work. They are complex to understand, or difficult to setup and use, and you may spend an enormous amount of time fighting with the program rather than doing the actual analysis. But it doesn’t have to be this way. dataGecko uses an award winning program called QlikView which takes away the pain of data analysis. We load your data into a QlikView report file and then custom design a set of filters, tables and graphs to suit that data. In a couple of days you have the data back in a standalone file that you can view on any PC running the QlikView application. You don’t have to be connected to any service, or even the net. If you have a laptop you can do your analysis anywhere you like, take it with you while you travel and show the results to colleagues in an instant. There are no costly servers, and no complex applications to worry about, it’s easier than using Excel, but far more powerful. You can use it to do live data investigation during meetings, or dynamic presentations at conferences. Once you are familiar with your report you will be astounded at the depth of information you can quickly dig up. And this is the key to the program: it’s the most powerful, flexible, yet simple to use data investigation tool on the market. And it’s easy to get this information out again as well. As you discover useful views of the data you can bookmark them so you can instantly return to that view in the future. You can export views of the data as text or graphics, or export directly to Excel. Creating research papers has never been easier as you can just grab the tables or graphs directly from the application to paste into your document.
And it is just so easy to use. Most people who use the program can be taught the basics in less than 5 minutes. After a day or two you will wonder how you ever got by without it.
What data can I analyse with this solution then?
Well, pretty much any data can be loaded into the application. The above examples talk about surveys, but really almost any data can be analysed this way. It is particularly good for large flat files (such as log file analysis), financial data or historical trend data, but it can load in data from a vast array of sources. We have used it to analyse data and reports from: phone call management systems, virus tracking databases, staff time-sheeting, internet terminal usage, web server usage, library OPAC usage, keyword searches, business management systems, asset management systems, KPI management, fault docket tracking systems, training booking systems, online banking transactions, and of course, online surveys. Surveys present special analysis challenges as the data can be complex in structure, especially for questions which provide multiple answers. dataGecko has developed special code to convert this tricky data into a format suitable for multi-dimensional analysis.
So what does dataGecko do exactly then?
We believe that the best person to analyse your data is you. You know the specifics of your field of research, the user population and the questions you are seeking to answer. Up till now though this has meant that you either had to become an expert in a complex analysis tool, which is a big investment in time and energy, or you had to rely of someone else’s interpretation of the data. With our solution you can take back control of the intellectual process of data analysis and investigation. The tool is so quick and easy to use that even a CEO can use it!
What dataGecko does is to take the data you provide, load it into the system via some clever scripting, and produce a number of logical data views to allow some initial data investigation. Then, working with you, we develop a set of customised tables and charts to allow detailed analysis based on your specific requirements.
Our suggested approach is:
- If possible, talk to us first about your project and we can suggest a course of action to make the data capture as effective as possible. If you want to analyse a survey, provide us with your questions and we can advise and make suggestions to maximise the outcome. Asking a few critical questions the right way can make all the difference to useful information, and just more useless data.
- If you are analysing an existing system or set of data, provide a good sample for initial investigation. We will review the data and provide a quote for the first stage of work.
- We will create an initial file that is very open ended and flexible that will allow you to undertake the initial data investigation, and get to know what’s under the “icing of the cake”.
- Once this is done we can provide a quote for the second stage of the work to provide a detailed, customised analysis report. We will work with you to create a set of specific tables and charts that will allow you to undertake further investigation and detailed analysis of the data.
How do you work with us?
dataGecko is a virtual beast, and we use the internet as our medium for communications. We can work with clients anywhere in the world. There is very little we cannot achieve remotely.
- Talk to us about your project:
Obviously the first step is to have a talk about your project, what you need to achieve and your budget. You can contact us via email to start with, and then we can talk in more detail over the phone. If it’s possible we may arrange a physical meeting, but it’s usually costly and unnecessary. This stage may also involve an analysis of your source data if you already have some available. We can then produce a quote for the first phase of the work. If our solution sounds suitable we will establish a private online project management space (using Google Groups) that will allow us to: define the project scope, requirements and timelines, discuss the project aspects in detail, collaborate on refining the content (wiki style), share project resources (including links/files/etc), and track project time and costs.
- Begin the project: Phase 1.
Once everyone is happy that the project is understood we can move to the next stage. This may include any or all of the following: research, analysis of data collection requirements, refinement of survey instruments, analysis of sample data, setting up and running online surveys, setting up the basic level QlikView report including the loading of data, and assistance with data interpretation. The basic level report will have every element of the data presented as simple tables of data, with filters created on all elements to allow complete data investigation. It will not have any customised views, tables or charts at this stage, but is sufficient to get to know your data intimately. You may want to spend a little time getting cosy with this data before moving to phase two.
- Review Phase 1 and continue to Phase 2.
Phase one will have generated a completed basic level QlikView report file which you can use for initial data investigation. It may be that this is all that you require, or you may wish to customise the report further to provide specific report views, tables and charts. Again we will provide a quote to undertake this work based on your requirements. This will produce a final report customised to your requirements that can be used to generate data for written reports, or as a dynamic presentation tool in its own right.
To assist with working remotely we use a number of tools. Our own web site will have a project entry with links to the various tools for your project. Google Groups provide a great project management tool with an excellent collaborative authoring system, and we use this service extensively. For online surveys we suggest SurveyMonkey if you want to manage the survey yourself, or we can create and host a customised survey using our own survey system. Our toolsets provide you with access to your project 24×7 from anywhere in the world.
What does it cost?
This is of course dependent on many factors, but we realise that keeping costs to a minimum is critical for everyone these days. As such we break the costs down into three components: QlikView licencing, fixed costs and hourly rates.
QlikView licencing: The QlikView analysis program is available in various versions. To begin with there is a free trial version available that will enable you to review the product and even try it with your sample data. It is time restricted, but with judicial timing you can see what your final report will look like using just the free trial licence. Beyond that you will need the basic Analyser version of the product which provides unlimited use of this and any future reports. If the source data format remains fixed, you can even load in updated versions of the raw data in the future, to produce say an ongoing monthly report from the same data source. Each person who needs to undertake analysis of the data requires a licence, but remember that result data can easily be exported to Excel/Word for wider distribution as required. If you want more control you can purchase the Professional licence which provides the additional power to add/modify/delete report tables, graphs and other visual elements to the report. There is also the ultimate Enterprise level licence, which provides complete development control and allows unlimited new reports to be developed by your own staff. If your requirements grow large enough you can bring all the development work in-house. This provides peace of mind as you are not locked into our services in the future if that is your preference. We can assist with consulting and training on making this move, but obviously the costs are greater. With QlikView you can start off very small, and grow to enterprise level implementation should you wish. It is a very powerful, capable and flexible service.
Fixed Costs: These usually centre around development of say an online survey or a QlikView report. In most cases we can work out how much effort is required to produce a given report and quote for that up front. Coding of a report is dependent on size and complexity of the elements, but its usually very cost effective.
Hourly Rate: For all other tasks that are difficult to quote as a fixed costs we provide an hourly rate based quote. We might suggest that a given project requires say 6 hours to complete a given element, and we will track the time spent in the online management system. If you want more done beyond that, we will quote again for additional hours. This is useful, for example, if you would like some analysis of the data done on your behalf. You can fix a budget of say 4 hours of analysis time, which we will do and provide you with the results of that. Should you want more, you simply buy more time. We will work closely with you to make the process as cost effective as possible.
As a guide you should allow around $1000 for the QlikView Analyser licence, which is a one off cost, and then between $500 and $1000 for an average sized QlikView Survey Report, depending on how much extra hourly rate work is involved. This cost is quickly recovered in time savings when doing your analysis. For more detailed consulting projects these costs will be more. For organisations this represents a very cost effective solution.
Assuming you have created your own survey questions, setting up an online survey cost between $200 and $500 on average including hosting, or roughly around $100 per 10 questions, and data cleansing is based on an hourly rate of $30-$50 depending on the data. (Or you can do your own surveys on SurveyMonkey and save money here, and just get us to create the QlikView report.)
Of course, actual projects will be quoted based on the specifics of each project, so the above figures are only a rough guide. In all cases you should allow some additional time for the specification process and possibly for results interpretation.