I think Kodaira did important work on almost complex structures, but might be wrong.
Bourbaki exp 35.
nLab page on Almost complex variety