L. Petrov, "Using source maps for scheduling and data analysis: approaches and strategies", Proceedints of the 18th European VLBI for geodesy and astrometery working meeting, ed. by J. Boem, A. Pany and H. Schuh, p. 141-146, 2007. ABSTRACT Since almost all the sources observed with VLBI at intercontinental baselines are resolved, information about source structure should be utilized during scheduling and analysis of geodetic VLBI experiments. I collected the database of images of $\sim 4000$ sources from analysis of VLBA and Australian LBA array observations. Analysis of images and calibrated visibilities allowed us to select a list of 230 of the best sources for including in geodetic observing schedules. These sources have a)~a significant correlated flux density at long baselines, b)~a significant compactness index, and c)~a least asymmetric core. For these sources the coherence function depends mainly on the length of the projected baseline, and neglecting its dependence on baseline orientation produces errors in the coherence amplitude prediction less than 20\%. Usage of source maps in the form of an expansion over a set of $\delta$-functions in data reduction process requires great care. I found that the traditional hybrid CLEAN algorithm produces maps which are not always suitable for reduction of geodetic experiments, since the noise in data makes computation of the contribution of source structure to group delay highly instable.