I would like to know how possible it is, to locate the depth of a magma chamber from where a lava flow is issued when you have its chemical composition .
From whole-rock geochemical data ALONE I would have thought that might be quite difficult, but if you combined this with the mineral chemistry of phenocryst phases, you might be able to derive some P-T data. P-T data for magmatic rocks are typically derived from the experimental determination of the partitioning of trace elements between crystals (phenocrysts) and magma (whole-rock), and then using those models for real-world examples. However, you must also have some ideas as to whether the phenocryst population crystallized in-situ or whether they are cumulative or mixed-in, so some petrographic examination is also required.
Well the geothermometry/geobarometry literatue is vast, and pointing you at a few papers that will actually help is actually quite a challenge. Determining the P/T conditions of crystallization are obviously in part dependent on the composition of the magma, so are we talking about basalts, andesites or rhyolites here? Also, the tectonic setting will have huge implications for the geothermal gradient involved and the overall P/T conditions you might expect. So to be honest, you will probably be better served doing your own literature search using terms such as ‘geobarometry’, ‘basalt’, ‘island arc’, substituting the relevant terms for your particular geological environment.