• about us

  	humangraph our philosophy range of activities experience

• research methods

  	quantitative research qualitative research Mystery Shopping

• research areas

  	strategy and development of new products sales, distribution and trade marketing Category Management effectiveness of promotional and marketing activities CRM HR research

• eye tracking

  	visual track shopper track

• data analisys

  	data modelling reporting

• RFP

  	telephone contact send us your brief complete our brief form

• reports from research
• keep smiling
• contact

data modelling

The philosophy of HumanGraph is to treat each customer in an individual way, taking into account their research experience, industries in which they operate, competitive environment, type of Customers etc.. Therefore:

we do not offer our Customers ready research models or standardized models of questions (without the possibility of changing their contents ...)

Thanks to effective analytical tools and cooperation with researchers specializing in statistics and quantitative methods of data analysis, we offer our customers a wide array of simple and advanced analyses of the following types:

Simple correlations and regressions (e.g. linear) - we show the relation between individual variables, the strength of such a relation, and we try to predict how a change in one element may affect another element.

Conjoint analysis (determining purchase preferences of customers and finding an optimal profile of services/product, determining the importance of individual attributes and their influence on shopping patterns)

Logistic regression (additionally specifies the probability of occurrence of certain behaviours/events, ideal in case of segmentation of databases and predicting reactions to an offer)

Correspondence analysis (perception maps, image, features of an ideal product / service)

Latent class analysis (predictive character, analysis of probability of an entity belonging to each defined segment)

Structural equations (allow you to specify the force of influence of independent variables on the variable being explained...)

Kohonen networks, also called SOM (these project complicated "real" relations between respondents, grouping them in natural segments)