Data Mashups vs. OLAP Cubes

Eric Kavanagh (EK): That’s very interesting. And Malcolm, I’ll bring you in, too, because again sometimes we get kind of lost in definitions and we’re splitting hairs essentially but it seems to me the whole concept of a mashup is that you really want to enable a very swift and agile world of analysis for your end users, and some of the dashboard products out there are getting a lot easier to use.

There are data visualization technologies that are a lot easier to use than they were three-four years ago. So it all seems to be moving in that direction of empowering the end user to mix and match data sets. But it seems to me that ideally the beauty of a mashup environment, if it is done properly, is that you can essentially, I don’t want to say circumvent IT, but you can avoid a lot of the painstaking work required for building specific OLAP cubes. Is that a fair assessment?

Malcolm Chisholm (MC): I think so. I think that it’s certainly enables the end users to act in an agile way. But for me, I also think that situational awareness is also important which is, if I come in the morning, and I have a really good dashboard I can look at my production data landscape and just see very quickly that everything is OK. Or I can look at the previous example of the UN country mashups and just see what’s going on.

Getting Situational Awareness In A Rapid Manner

I don’t necessarily have to act. But if I was to try and gain that knowledge by reading reports, it would take me forever. I really wouldn’t have the time to do it. I wouldn’t be able to get there. So getting situational awareness in a sort of rapid manner that is easy to assimilate, I think is as critical as the ability to act in agile fashion.

EK: That’s a good point. William, one last comment from you before the first break. Is there some reason in your mind why, for example, the stuff like Salesforce and this other, almost like, SMB type stuff seems to be picking up on this faster than the enterprise?

I know of, for example, some really high profile mashup environments at places like USAA which is obviously a huge financial services company. But by and large, you don’t hear it as much from the larger institutions. Is there any reason why you think that might be the case?

William Laurent (WL): Yeah, because the primary factor in my opinion is this, that well over 50% of mashups in existence are based on geographical information. So you have your first responders that in an emergency deportment, fire department.

Getting back to what Malcolm was talking about, the situational awareness, you have a command system that is in place to respond to various events on the ground, and those types of things, tracking sales, these type of applications are what’s out there right now, and the geographic paradigm really is fundamental to mashups. That’s really what has driven this first generation of data mashups.

What Are the Advantages of Data Mashup Over OLAP Cubes?

Data mashup and OLAP (Online Analytical Processing) cubes both serve the purpose of integrating and analyzing data, but they do so in different ways, each with distinct advantages and trade-offs. While OLAP cubes were a popular tool in traditional business intelligence environments, data mashup tools have become a preferred option for modern, agile data analysis. Here's a detailed comparison of the advantages of data mashup over OLAP cubes:

1. Real-Time Data Integration

• Data Mashup:
  One of the most significant advantages of data mashup tools is their ability to pull data from multiple sources and integrate it in real-time or near real-time. This means that users can see the most up-to-date information and make timely decisions without waiting for periodic data loads.
• OLAP Cubes:
  OLAP cubes are often based on pre-aggregated, static data that is refreshed on a scheduled basis. Changes in source data require reprocessing or reloading of the cube, leading to delays in getting updated information.

Advantage: Data mashup offers real-time or near real-time data analysis, whereas OLAP cubes are more static and require time-consuming data processing to reflect changes.

2. Flexibility in Data Sources

• Data Mashup:
  Data mashup tools are designed to pull data from a wide variety of sources—cloud-based platforms, databases, APIs, spreadsheets, social media, and even unstructured data. This flexibility allows for the integration of diverse data sets, both structured and unstructured, without needing to adhere to strict schemas.
• OLAP Cubes:
  OLAP cubes typically rely on structured, relational data sources and require predefined schemas. Adding new data sources or changing the cube's structure can be complex, requiring the involvement of IT teams and reengineering of the cube.

Advantage: Data mashup tools offer far greater flexibility in terms of the types and variety of data sources they can integrate, while OLAP cubes are more limited to structured data.

3. Faster and Easier Implementation

• Data Mashup:
  Data mashup tools are often more agile and easier to implement. They usually have a user-friendly, drag-and-drop interface that allows even non-technical users to integrate data and create reports. The implementation process can be quick, with minimal setup and configuration.
• OLAP Cubes:
  Building and maintaining OLAP cubes typically require significant time and effort. The development process involves creating a cube schema, defining dimensions and measures, and optimizing performance. This setup is usually done by database administrators or IT professionals and can take weeks or months.

Advantage: Data mashup tools can be implemented and used quickly by business users without heavy reliance on IT, while OLAP cubes require more time and specialized skills for setup.

4. No Predefined Schema or Modeling Requirements

• Data Mashup:
  Data mashup allows for on-the-fly data combination without the need for predefined data models. Users can pull data from various sources, combine it as needed, and explore it dynamically without worrying about upfront schema design. This is especially useful for organizations with diverse and changing data needs.
• OLAP Cubes:
  OLAP cubes require careful upfront planning and modeling, including defining dimensions, hierarchies, and measures. Changes in the data model can be time-consuming and often require reprocessing the cube, making it less flexible in dynamic environments.

Advantage: Data mashup does not require predefined schema design, allowing for more agility and responsiveness to changing data needs compared to the rigid structure of OLAP cubes.

5. Cost Efficiency

• Data Mashup:
  Data mashup tools are often more cost-efficient, particularly in environments where real-time data integration, cloud-based architectures, and flexibility are required. They eliminate the need for large data warehouses and reduce the dependency on ETL processes, which can drive up costs in traditional BI environments.
• OLAP Cubes:
  OLAP systems are usually part of larger, more complex data warehousing architectures, which can be expensive to maintain. The need for regular cube processing, specialized software, and technical expertise adds to the overall cost.

Advantage: Data mashup tools tend to be more cost-effective, especially for organizations looking for flexibility without the high infrastructure and maintenance costs associated with OLAP cubes.

6. Ad-Hoc Analysis and Self-Service BI

• Data Mashup:
  With data mashup tools, business users have more freedom to explore data without needing IT assistance. They can easily combine data from multiple sources, run ad-hoc queries, and generate insights without waiting for pre-built reports or cubes. This promotes a self-service BI environment, where users can independently analyze and visualize data.
• OLAP Cubes:
  OLAP cubes are typically designed with specific, pre-defined questions in mind. While they are optimized for fast querying within defined dimensions, performing ad-hoc analysis outside the cube's structure can be difficult. Non-technical users often need to rely on IT to make changes or create new cubes to answer different questions.

Advantage: Data mashup tools empower business users to perform ad-hoc analysis more easily, fostering a self-service BI culture, while OLAP cubes are more rigid and often require IT intervention for non-standard queries.

7. Scalability and Cloud Integration

• Data Mashup:
  Many data mashup tools are designed to work in cloud-based environments, making them easily scalable as data volumes grow. They can connect to cloud-based data sources and perform analytics in distributed environments, which is increasingly important for organizations leveraging cloud infrastructure.
• OLAP Cubes:
  OLAP cubes are generally more difficult to scale, particularly as data volumes increase. Expanding the cube often requires more resources, re-architecting the data warehouse, and optimizing performance. Traditional OLAP systems are also more likely to be on-premises, which limits their ability to take advantage of cloud-native scalability.

Advantage: Data mashup tools offer better scalability, especially in cloud-based architectures, compared to OLAP cubes, which often struggle with growing data volumes.

8. Support for Unstructured and Semi-Structured Data

• Data Mashup:
  Data mashup tools can handle both structured and unstructured data (such as social media feeds, text, logs, etc.), making them versatile for modern data environments where non-traditional data sources are increasingly important for business insights.
• OLAP Cubes:
  OLAP cubes are designed primarily for structured data in relational databases. While they excel at summarizing structured data, they are not built to handle unstructured or semi-structured data, limiting their usefulness in more modern, big-data contexts.

Advantage: Data mashup tools provide the ability to work with unstructured and semi-structured data, while OLAP cubes are limited to structured, relational data.

Previous: Data Mashups and Governance

Next: Implications of Mashup Technology for Enterprises