Data Mining Project Assessment
Data Mining Project Assessment
Successful data mining (also referred to as predictive modeling and business analytics) requires a purposeful blend of strategy and tactics. In the 1990s, pioneering companies realized the potential advantages of employing data mining technology as early as possible. They chose to undertake this initiative in-house. Too often, these innovators would prematurely dismiss the technology when they fell victim to common, yet elusive data mining project pitfalls. Other progressive organizations decided to maintain focus on their core competencies and outsource data mining. They too rarely met with success. Although their consultants may have had adequate analytics and industry experience, critical project management issues specific to data mining and their customers’ operations were overlooked.
Data mining projects are highly susceptible to failure without a comprehensive assessment and resulting plan. Many companies have engaged with large consulting firms to undertake major data mining projects only to find well into the project that the initiative should never have been started. Alternatively, a thorough assessment with applied data mining project management principles provides a flexible yet structured framework that anticipates and adapts for redirection from what is essentially a discovery process.
Assessment And Strategy Is Critical To Success
TMA follows data mining project management guidelines to provide accurate estimates of feasibility and effort, define responsibilities of the client-vendor team, and yet provide reasonable flexibility specific to the knowledge discovery process. One of the most critical components of this process is the Data Mining Project Assessment (DMPA).
The DMPA creates an essential foundation for a successful data mining initiative through understanding key stakeholder motivations and vision, leveling expectations, citing capabilities and limitations, evaluating all available resources, gaining an appreciation of each unit’s business processes, assuring end user adoption in advance, defining and valuating benchmarks for project performance, and requesting domain expert contribution for results interpretation.
The DMPA findings report will convey whether your organization is fully ready to proceed with business analytics, or document what is required to reach the starting line. The savings alone from not having embarked prematurely on a data mining project make the DMPA a substantial value. The findings and recommendations reports will provide situational feedback, near-term actionable steps and an overarching strategy adapted to a dynamic discovery process.
Preparatory Teleconference for introductions, overview, and to establish a meeting schedule for the on-site assessment.
On-site interviews (two days) with stakeholders, functional and database managers, domain experts and user group representatives to understand needs, objectives and resources. The interviews will generally follow the outline of the first stage of the “CRoss-Industry Standard Process for Data Mining” (CRISP-DM) —
- Determine Business Objectives
- Assess Situation, Environment and Resources
- Define Data Mining Goals
- Produce Project Plan
Topics to be investigated, analyzed and reported in the context of data mining feasibility applied to your application will include but not be limited to:
- Gain a high-level understanding of the business situation and determine what questions should be answered through data mining along with motivations of all members
- Start with the end in mind: how should a predictive model function from both strategic (decision support) and tactical (IT and user) perspectives?
- Understand the current processes in place relative to decision support
- Catalogue available resources: human, informational, structural, intra- and extra-organizational
- Establish a risk level and approach for the model: what is an appropriate tolerance for false positives or negatives?
- Identify how data is captured, stored and reported
- Visually inspect the data sources, structure and attributes
- Determine where check points may be established to monitor model performance and establish a feedback loop
- Set realistic expectations for data mining with respect to your organization’s business situation and resources
- Verify that data mining capabilities and limitations are well aligned with objectives: accuracy versus explainability will drive method selection
- Define metrics and benchmarks, translate to dollar value, map back to stated objectives and close the process loop
Assessment Briefing delivered within four weeks of the on-site visit through a PowerPoint presentation via web conference, or on-site meeting at an additional cost.
Assessment Notes will be a collection of written findings, plans, and other artifacts amassed during the course of the DMPA, delivered within the PowerPoint file which will become the property of your organization.
Actionable Recommendations will be presented in the form of a proposal, representing a customized forward-looking plan with clearly defined next steps that your organization may proceed with independently, or with TMA’s continued guidance. The document will become the property of your organization under the terms of a confidentiality agreement.
TMA’s DMPA is a fixed-cost service. Schedule a no-obligation discovery call to obtain current pricing and explore whether a DMPA makes sense for your situation. Extend an inquiry.